The Eastern Enlargement of the Eurozone

THE EASTERN ENLARGEMENT OF THE EUROZONE

The Eastern Enlargement of the Eurozone Edited by

MAREK DABROWSKI Center for Social and Economic Research, Warsaw, Poland and

JACEK ROSTOWSKI Center for Social and Economic Research, Warsaw & Central European University, Budapest, Hungary

A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN-10 ISBN-13 ISBN-10 ISBN-13

0-387-25764-0 (HB) 978-0-387-25764-8 (HB) 0-387-25766-7 ( e-book) 978-0-387-25766-2 (e-book)

Published by Springer, P.O. Box 17, 3300 AA Dordrecht, The Netherlands. www.springeronline.com

Printed on acid-free paper

All Rights Reserved © 2006 Springer No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed in the Netherlands.

TABLE OF CONTENTS

List of Tables ................................................................................................vii List of Figures................................................................................................ix Contributors ...................................................................................................xi Acknowledgements.......................................................................................xv Chapter 1. When Should the New Member States Join EMU? ......................1 J. Rostowski

Chapter 2. The Exchange Rate: Shock Generator or Shock Absorber?........15 M. Maliszewska, W. Maliszewska

Chapter 3. Do the New Member States Fit the Optimum-Currency-Area Criteria? ........................................................................................................41 M. Blaszkiewicz-Schwartzman, P. Wozniak

Chapter 4. EMU Enlargment and Trade Creation ........................................63 M. Maliszewska

Chapter 5. Future EMU Membership and Wage Flexibility.........................75 A. Radziwill, M. Walewski

Chapter 6. Exchange Rate Regimes and Nominal Convergence ..................91 M. Szczurek

Chapter 7. EMU Enlargement and the Choice of Euro Conversion Rates .113 L. W. Rawdanowicz

Chapter 8. The Short-Run Macroeconomic Effects of Discretionary Fiscal Policy Changes.................................................................................131 J. Siwinska, P. Bujak

Chapter 9. How to Reform the Stability and Growth Pact .........................147 J. Rostowski

Chapter 10. Uneven Growth in a Monetary Union.....................................159 N. Zoubanov

Chapter 11. ECB Decision-Making in an Enlarged EMU ..........................183 W. Paczynski

Chapter 12. A Strategy for EMU Enlargement...........................................199 M. Dabrowski

Index ...........................................................................................................227

v

LIST OF TABLES Table 1.1. Degree of trade integration of NMS with EMU compared to that of EMU countries............................................................................................................... 3 Table 2.1. Average inflation rates in countries grouped according to the RR classification ...................................................................................................... 29 Table 2.2. Average annual GDP growth in countries grouped according to the RR classification ...................................................................................................... 30 Table 2.3. Inflation and growth performance according to the ‘coarse’ RR classification ...................................................................................................... 31 Table 2.4. Level of inflation: ‘official’ (GGW) classification with various samples 32 Table 2.5. Level of inflation: the ‘natural’ (RR) classification with various samples ............................................................................................................. 33 Table 2.6. Level of inflation: the ‘official’ (GGW) classification with money growth excluded............................................................................................................. 34 Table 2.7. Level of inflation: the ‘natural’ (RR) classification with money growth excluded............................................................................................................. 35 Table 2.8. First-differences of inflation: the ‘natural’ (RR) and ‘official’ (GGW) classifications .................................................................................................... 36 Table 2.9. GDP growth per capita: the ‘official’ (GGW) classification ................... 37 Table 2.10. GDP growth per capita: the ‘natural’ (RR) classification...................... 38 Table 2.11. Volatility of GDP growth per capita (3-year SD): the ‘official’ (GGW) classification ...................................................................................................... 39 Table 2.12. Volatility of GDP growth per capita (3-year SD):the ‘natural’ (RR) classification ...................................................................................................... 40 Table 3.1. Trade with EU-15 as % of GDP, 1999-2003 ........................................... 43 Table 3.2. Trade with EU-15 as % of GDP, 1992-1995 ........................................... 44 Table 3.3 Trade with OMS as % of total trade, 1999-2002 ...................................... 45 Table 3.4. Correlation between annual real growth rates in the Eurozone and selected countries............................................................................................... 47 Table 3.5 Correlation of business cycles at quarterly frequency .............................. 49 Table 3.6. NER volatility: distance from the Club Med average.............................. 51 Table 3.7. Short-run (monthly data) volatility (NMS) .............................................. 55 Table 3.8. Short-run (monthly data) volatility (member states)................................ 56 Table 3.9. Long-run (quarterly data normalized to monthly) volatility (NMS)........ 57 Table 3.10. Long-run (quarterly data normalized to monthly) volatility (member states) ................................................................................................................. 57 Table 4.1. Trade effects of EMU............................................................................... 65 Table 4.2. Trade effects of EMU............................................................................... 66 Table 4.3. Results of the estimation of the bilateral gravity equation....................... 69 Table 4.4. Estimated potential trade flows between EU-15 and NMS. .................... 70 Table 4.5. Trade with the EU-15 as a share of GDP at current prices in 2002 ......... 70 Table 4.6. Estimated potential trade flows between NMS. ....................................... 72 Table 5.1. Explaining the dynamics of DULC.......................................................... 84 Table 5.2. The correlation between nominal and real exchange rate depreciation*.. 84

vii

viii Table 7.1. Exchange rate regimes in the NMS........................................................ 115 Table 7.2. Correlation coefficients between the Euro RER and relative prices...... 121 Table 7.3. PPP exchange rates, 1999 (national currency per Euro) ........................ 123 Table 8.1. Review of the empirical literature on non-Keynesian effects................ 143 Table 8.2. Estimation results; OECD countries; dependent variable: 'cit .............. 145 Table 8.3. Estimation results for transition economies; dependent variable: 'cit ... 145 Table 9.1. Small is Beautiful (1997-2002) .............................................................. 151 Table 11.1. The Pros and Cons of various ECB reform options............................. 191 Table 11.2. Estimated allocation of MS to groups (according to the methodology adopted by the EU Council and the ranking based on shares in ECB capital)195

LIST OF FIGURES

Figure 5.1. GDP real growth rates............................................................................. 87 Figure 5.2. Unemployment rates ............................................................................... 87 Figure 5.3. DULC Growth rates ................................................................................ 88 Figure 5.4. Productivity growth rates ........................................................................ 88 Figure 5.5. Real wage growth rates ........................................................................... 89 Figure 5.6. Nominal wage growth rates .................................................................... 89 Figure 5.7. Real depreciation rates (against the Euro) .............................................. 90 Figure 5.8. Inflation rates .......................................................................................... 90 Figure 6.1. Target exchange rate with limited (left) and high (right) reserves ......... 96 Figure 6.2. Exchange rate and fundamentals on the way from ERM-II to EMU ..... 99 Figure 6.3. State (left) and time and state (right) dependent regime switches........ 104 Figure 6.4. Surprise (left) and pre-announced (right) switches into a currency band.................................................................................................................. 107 Figure 6.5. Euro share in the optimal basket........................................................... 108 Figure 10.1. The impact on output of BG of an increase in demand for it in S...... 173 Figure 11.1. Median inflation rates (HICP) in EMU-12, (hypothetical) EMU-25 and in the four largest EMU-25 economies, January 1998- September 2004 (annual change, % points . .............................................................. 187 Figure 11.2. Median inflation in the EMU-12 compared to EMU12 inflation, January 1998- September 2004 (annual change, % points) ............................ 187 Figure 11.3. Median inflation in a hypothetical EMU -25 compared to EMU-25 inflation, January 1998- September 2004 (annual change, % points)............. 188 Figure 11.4. The size of the EMU and the distribution of voting rights ................. 193

ix

CONTRIBUTORS

Marek Dabrowski is Professor of Economics, Chairman of the CASE Foundation Council, and was First Deputy Minister of Finance of Poland (1989-1990) and Member of the Monetary Policy Council of the National Bank of Poland (19982004). In the last decade he has been involved in policy advice and policy research in Belarus, Bosnia and Herzegovina, Bulgaria, Georgia, Iraq, Kazakhstan, Kyrgyzstan, Macedonia, Moldova, Mongolia, Romania, Russia, Serbia, Turkmenistan, Ukraine and Uzbekistan and in a number of international research projects related to monetary and fiscal policies, currency crises, EU and EMU enlargement, and the political economy of transition. Jacek Rostowski is Professor of Economics at the Central European University, Budapest, Adviser to the CEO of Bank Pekao, Warsaw, a Member of the CASE Foundation Council, Senior Fellow at CASE, and an Associate of the International Economics Program at the Royal Institute of International Affairs, London. During 1997-2001 he was Chairman of the Macroeconomic Policy Council at the Polish Ministry of Finance and during 2002-4 he was Adviser to the Governor of the National Bank of Poland. He has also advised on the transition from communism in Belarus, Latvia, Russia, Serbia and Ukraine. He has published widely on monetary and financial systems and on the macroeconomics of transition. Monika Blaszkiewicz-Schwartzman received an MA in International Economics from the University of Sussex (2000) and MA in Economics from the University of Warsaw (2001). Between 2000 and 2002 she worked for the Ministry of Finance (Poland). Since October 2002 she has been a member of the Maynooth Finance Research Group (MFRG) affiliated to the Institute of International Integration Studies (IIIS), based at Trinity College Dublin (Ireland) working on her PhD thesis. Her main field is international macroeconomics. Research interests include financial crises and European integration (EMU). She has collaborated with CASE since 2000. Piotr Bujak has collaborated with CASE since 1999. Among others, he has coauthored CASE’s quarterly Polish Economic Outlook and participated in advisory projects in Georgia (2000-2002) and Serbia and Montenegro (2001). His main areas of interest include monetary policy, fiscal policy and the economics of pensions. Currently he is an economist at Bank Zachodni WBK SA dealing with the Polish macro-economy and fixed income analysis. Maryla Maliszewska graduated from the University of Sussex (1996) and Warsaw University's Department of Economics (1997). She received her PhD from the University of Sussex in 2004. She has been working with CASE since 1996. Her research interests cover modeling of international trade flows, determinants of real exchange rates, and location of production and agglomeration externalities in transition. Her study on the impact of Poland's accession to the EU within a xi

xii computable general equilibrium framework was awarded second prize at the annual Global Development Network Research Medals Competition for ‘Outstanding Research in Development’ in January 2004. Between 1997-98 and in 1999, she worked as CASE’s representative at the ProDemocratia advisory mission in Romania. Wojciech Maliszewski graduated from the University of Sussex (1996), Warsaw University's Department of Economics (1997) and the London School of Economics (1999). He is reading for his doctorate at the London School of Economics. He has been associated with CASE since 1996. His research interests cover central bank independence, exchange rate behavior, monetary policy transmission mechanisms and fiscal policy in transition economies. During 1997-99 he worked in Romania as CASE’s representative at the ProDemocratia advisory mission. Recently he has joined the International Monetary Fund where he is an economist in the Fiscal Affairs Department. Wojciech Paczynski has been an economist at CASE since 2000. His research interests include applied macroeconomics, game theory and political economy. He holds Master degrees in International Economics from the University of Sussex (1998), Warsaw University's Department of Economics (1999) and Warsaw University's Department of Mathematics and Computer Science (2000) and is currently reading for his PhD at the University of Dortmund. Between 2001 and 2004 he was the editor of The Global Economy, a CASE quarterly analytical/forecasting publication. Artur Radziwill, is Vice President of CASE and has been professionally interested in post-communist transition and economic development for many years. He has extensive practical experience in these areas including technical assistance, consultancy and research projects with the World Bank, UNDP, ILO, GTZ and the Open Society Institute. He headed the team that prepared the country study for Moldova in the GDN “Explaining Growth” project. Currently a Ph.D. candidate at the Department of Economics at Birkbeck College, London, he obtained his MA Summa Cum Laude in Economics from Warsaw University in 1998. He has also studied at the University of Sussex (MA International Economics). Lukasz Rawdanowicz graduated from Sussex University in 1998 (MA in International Economics) and Warsaw University, Department of Economics in 1999 (MA in Quantitative Methods). He is a co-editor of the Polish Economic Outlook and “Global Economy”. He was also involved in the development of a macro-model of the Kyrgyz Republic (2000). In 2003 he worked for the Economic Department of the OECD and since 2004 he has worked for the European Central Bank. His main area of interest is international macroeconomics, trade, econometric analysis and macroeconomic forecasting.

xiii Joanna Siwinska has collaborated with CASE since 1997.She holds a Ph.D. from Warsaw University (2002). She graduated in 1996 from the University of Sussex (UK) and the Department of Economics at Warsaw University (1997). As a Fulbright Scholar, she was a visiting researcher at Columbia University, in New York, USA (2000/2001). She specializes in fiscal policy, public finance and growth theory. Mateusz Szczurek is the chief economist at ING Bank in Poland, covering macroeconomics and fixed income analysis. He is finishing his PhD at the University of Sussex. His doctoral thesis is on international liquidity and foreign exchange crises. He holds an MA in International Economics from the University of Sussex, an MA in economics from the University of Warsaw, as well as a BA (joint honors degree) from Columbia University and Warsaw University. He is a member of the Polish Society of Market Economists and of the Polish Association of Business Economists. He has collaborated with CASE since 1998. Mateusz Walewski graduated from the Department of Economics at Warsaw University (MA in 1998) and the University of Sussex in the UK (1997). He has worked for CASE since 1997, and has participated in numerous CASE research and advisory projects and is the author (or co-author) of several publications concerning labor markets, inflation, restructuring of the economy and taxation policy, which are his main areas of interest. He also writes for CASE’s quarterly Polish Economic Outlook (PEO) on labor market issues. During 2000-2002 he was resident expert in CASE’s advisory project for the President of Georgia and since 2004 he has been a member of the Board of Directors of CASE Transcaucasus in Tbilisi. Przemyslaw Wozniak obtained an MA (1997) and a PhD (2002) in Economics from Warsaw University. He also studied at the University of Arizona (1995-1996) and Georgetown University (1998-1999) in the United States. He specializes in issues related to inflation, core inflation and monetary policy. He has participated in numerous research projects in CASE and published several papers including studies on core inflation in Poland and Ukraine, as well as on relative prices and inflation in the CEE countries. His work experience includes internships at the World Bank in Washington DC in Montenegro as economist for the quarterly ‘Montenegro Economic Trends’. In 2001 he was a Visiting Scholar at the International Monetary Fund in Washington DC (GDN/IMF Visiting Scholar Program). He has collaborated with the CASE Foundation since 1996 and since 2004 he has been a Member of the CASE Foundation Council. Nikolai Zoubanov graduated in 2003 with an MA in Economics from the Central European University (Hungary), he is currently reading for a PhD at the University of Birmingham (UK). His research interests include: the economics of European integration and transition in Central and Eastern Europe and the political economy and economics of human migration. Membership in academic organizations: the European Policy Network (UK), the Network of Industrial Economists (UK).

ACKNOWLEDGEMENTS

This volume contains a part of the results of the research project on: ‘Strategies for Joining the EMU’, grant No. 2 H02C 029 23 of the State Committee for Scientific Research (of the Government of Poland) undertaken between November 2002 and November 2004 by CASE – the Center for Social and Economic Research in Warsaw. The project team consisted of Marek Dabrowski and Jacek Rostowski (project coordinators), Maágorzata Antczak (project secretary), Monika BlaszkiewiczSchwartzman, Michal Brzozowski, Piotr Bujak, Michal Gorzelak, Marek Jarocinski, Przemyslaw Kowalski, Maryla Maliszewska, Wojciech Maliszewski, Anna Orlik, Wojciech Paczynski, Artur Radziwill, Lukasz Rawdanowicz, Joanna Siwinska, Mateusz Szczurek, Mateusz Walewski, Przemyslaw Wozniak and Nikolai Zoubanov. Participants of the project prepared 20 research papers published preliminarily as working papers in the CASE “Studies and Analyses” series. This volume presents 12 comparative papers, which cover two broad groups of problems. The first group addresses the theoretical and practical issues related to the process of EMU accession itself, such as the implications of the theory of optimum currency areas, the degree of trade and economic integration of NMS with the EU and Eurozone, the progress in nominal and real convergence that has already been achieved, as well as the expected additional benefits for trade creation for NMS from adopting the Euro, the nature of shocks experienced by the NMS, their labor market rigidities and fiscal challenges and, finally, the dilemmas associated with the ERM-II mechanism and the choice of the euro conversion rate. The second group of papers deals with the functioning of an enlarged Eurozone and discusses key questions such as the role of counter-cyclical fiscal policy, the future of the Stability and Growth Pact (SGP), the voting mechanism in the ECB Governing Board and to what extent a single interest rate policy fits the economic realities of countries with very different rates of productivity growth. Original papers were revised and updated as necessary. Chapter 1 is a revised version of an article that first appeared in ‘International Affairs’ (London), vol. 79, no. 5, October 2003, and is reproduced with permission. Claire Wilkinson and Nikolai Zoubanov helped with the language editing, and Mykyta Mykhaylychenko with the technical editing of this volume. Marek Dabrowski and Jacek Rostowski February 28, 2005

xv

JACEK ROSTOWSKI

CHAPTER 1

WHEN SHOULD THE NEW MEMBER STATES JOIN EMU?

1. INTRODUCTION After the EU accession, the new member states (NMS) are actively considering when to join the EMU. Unlike Britain and Denmark the NMS will not have an ‘optout’ - the right to remain outside the EMU indefinitely. The question is therefore only one of timing. A number of countries have already declared that they want to join EMU as soon as possible after EU accession. The costs of such a move are minimal for Estonia and Lithuania, which already have their currencies ‘hardpegged’ to the euro through a currency board (Sulling 2002)1. On the other hand, the European Commission (Commision, 2002, 10 Oct.) and the Bundesbank (Bratkowski and Rostowski 2002, EU hopefuls, 2000, 26 Sept.) have also made clear their lack of enthusiasm for new members joining the EMU quickly. What are the reasons pushing many NMS towards early EMU entry, and why are some key players in Western Europe opposed to it? We divide the discussion into seven sections. Section 1.2 describes the macroeconomic risks to which NMS are exposed while they remain outside the EMU. Section 1.3 assesses the standard optimum currency area (OCA) criteria for whether these countries are ready for EMU accession. Section 1.4 considers whether there is a trade-off for NMS between joining EMU and real economic convergence with the EU. Section 1.5 addresses some arguments which claim that NMS accession to EMU would be bad for current EMU members. Section 1.6 considers institutional issues and Section 1.7 concludes. 2. RISKS OF MACROECONOMIC INSTABILITY OUTSIDE THE EMU In earlier papers we have analyzed the difficulties faced by the more advanced central and east European countries in achieving macroeconomic stability in the runup to EU and EMU membership (Rostowski 2000), and have presented the case for unilateral euroization as a solution to these problems (Bratkowski and Rostowski, 2000). At present, unilateral euroization seems unlikely to be adopted by any of NMS. Nevertheless, the problems we have identified remain, and in the absence of unilateral euroization, rapid adoption of the euro becomes the best remedy.

1 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 1-14. © 2006 Springer. Printed in the Netherlands.

2

JACEK ROSTOWSKI

The argument goes as follows: successful market reforms in NMS leads to expectations of rapid economic growth. This in turn means that domestic residents wish to save less so as to smooth consumption, while foreign investors are willing to provide the financing needed to bridge the gap between savings and investment. The result is high capital account surpluses and their corollary high current account (CA) deficits. This, in turn, makes the NMS very susceptible to capital inflow ‘stops’ (reversals are not necessary) possibly leading to currency crises. We also argue that neither monetary nor fiscal policies can be counted on to keep these developments in check. Under a floating exchange rate regime, contractionary monetary policy will cause the domestic currency to appreciate, which may increase the CA deficit even further. In fact, very contractionary monetary policy has led to a significant reduction in the current account deficit in Poland in 2001-2003. Our argument is not that this cannot happen, merely that it cannot be counted on to happen. Furthermore, the reduction in the CA deficit has occurred at a very high cost in foregone output (growth had to fall sharply so as to reduce the motive for consumption smoothing). Expansionary monetary policy will lead to faster inflation and is likely to make the achievement of the Maastricht criteria impossible. Under a fixed exchange rate, monetary policy is not available as an instrument. Turning to fiscal policy (which can be used with either floating or fixed exchange rate regimes), this may also prove ineffective in improving the CA. A tightening of the fiscal stance may simply make foreign lenders more willing to lend to domestic private sector borrowers (we know that foreign investors do nowadays look at the overall indebtedness of a country’s residents, both public and private)2. In any event, even assuming a very high (one to one) correspondence between fiscal and current account changes, reducing a current account deficit by three percentage points would involve a politically improbably large fiscal adjustment. Expansionary fiscal policy would, in the traditional way, increase aggregate demand and thus tend to increase the CA deficit. Given the difficulty of reducing high CA deficits, many NMS are very exposed to the risk of a sharp depreciation of their currency, commonly called a currency crisis. In countries with high levels of ‘liability euroization’ (Calvo, 1998) such crises will lead to increases in the real debt burden and to depression (Indonesia post-1997 is a recent example). Many NMS do not have a very high level of liability euroization (e.g. the Czech Republic, Hungary, Poland) but even in these the need to offset the inflationary effect of a sharp depreciation through higher interest rates could well lead to a strongly recessionary effect. In the absence of unilateral euroization, rapid accession to EMU is the only way of escaping the choice between slow growth and high exposure to the risk of currency crisis (possibly followed by a prolonged period of slow growth, as occurred in the Czech Republic after the 1997 crisis). Thus, the period before NMS have joined the EMU exposes these potentially fast growing countries to either a high risk of currency crisis, or forces them to grow far more slowly than they could. Since real convergence is one of the purposes of EU accession for the NMS, the orthodox path is at variance with the ultimate goal, something which cannot be desirable. This is the crux of my argument for early EMU accession for NMS.

WHEN SHOULD THE NEW MEMBER STATES JOIN EMU?

3

3. OPTIMUM CURRENCY AREA CONSIDERATIONS Asymmetric shocks are a danger if a country and the monetary union it proposes joining are not part of an OCA. Our view is that many NMS satisfy the OCA conditions to the same degree as present members of EMU, or are very close to doing so. Since the NMS are committed by their acceptance of the acquis communautaire to joining the EMU at some stage, what is good enough for the EMU's current members should be good enough for the NMS. Even if some of the NMS satisfy OCA conditions a little less than current EMU members, this merely exposes them to slightly higher risks from idiosyncratic shocks than current EMU members are exposed to. In making a choice on EMU accession, these slightly higher risks must be set against the very high costs described above of keeping one's national currency in the pre-accession period. Table 1.1. Degree of trade integration of NMS with EMU compared to that of EMU countries Country Belgium-Lux. Hungary Czech Republic Estonia Slovenia Slovakia Netherlands Ireland Bulgaria Portugal Austria Romania Latvia Poland Lithuania Spain France Germany Finland Italy Greece

EMU trade/GDP 81.4 73.2 65.1 62.0 61.8 58.9 48.8 44.2 39.3 38.5 37.6 34.7 30.9 27.6 26.5 25.5 21.7 20.8 20.7 19.5 17.4

Intra-industry trade/GDP(estimated)* 59 43 43 24 37 29 38 22 13 19 26 10 7 12 6 17 18 17 10 12 5

EMU

EMU trade/total trade 56.8 68.7 61.7 45.1 67.1 56.8 47.9 33.2 54.2 67.1 63.2 66.4 46.8 58.5 36.0 58.3 51.9 43.8 34.0 49.3 53.4

* The shares of intra-industry (II) trade with EMU countries in GDP were estimated by taking the 1997 shares of II trade with the EU in Fidrmuc and Schardax (2000) and applying them to columns 2 and 4. Source: Eurostat. Data is 1999 for NMS and 1998 for EMU countries.

The main reason we think that many NMS are close to satisfying OCA requirements to a similar degree as current EMU members is their very high level of trade integration with EMU countries. Trade with other members of a currency area as a share of GDP is a good indication of the extent to which idiosyncratic shocks to

4

JACEK ROSTOWSKI

a country's economy are likely to be amortized by its trade with the rest of the currency area. Table 1.1 shows that in 1999 all of the NMS traded a higher share of their GDP with EMU countries than six of the current 12 EMU members (including the four largest Germany, France, Italy and Spain) in the year preceding the launching of the euro3. Trade with other members of a currency area as a share of total trade is a good indication of the extent to which a country would be exposed to movements in the exchange rate of the common currency against the currencies of ‘third countries’: all of the NMS traded a higher share of their total trade with EMU countries than two EMU members, and six of the NMS traded a higher share of total trade with EMU than all but two current EMU members. Thus, if these ratios were the only criteria for satisfying OCA requirements, we could already conclude that many NMS satisfy them4. It is argued (e.g. Fidrmuc and Schardax, 2000) that a higher share of intraindustry (II) trade within a currency area will lead to more synchronous business cycles, because industry specific supply or demand shocks are then more likely to be symmetric across countries. We therefore measure II trade with EMU/GDP for EMU members and NMS, and find that seven NMS have a share of II trade with EMU/GDP which is higher than that of three EMU members. II trade is often classified as coming in one of two varieties. ‘Horizontal II trade’ where almost identical goods are exchanged between two trading partners and ‘vertical II trade’, where the goods traded although they come under the same IST classification, differ in some important respects. It is usually assumed that a given amount of ‘vertical II trade’ assures less correlation between business cycles than the same amount of ‘horizontal II trade’ of almost identical goods. This is not clear for two reasons: (1) if vertical trade depends on the supply of components by one side and of finished products by the other, then the two countries’ trade and output flows are likely to be very highly correlated indeed; (2) if vertical trade depends on exchanging similar products of differing quality, then if NMS export the lower quality, cheaper variants, the income elasticity of demand for these is likely to be lower than for that of the corresponding EMU country products. In that case, aggregate or sectoral demand shocks in EMU economies are likely to affect NMS less (falls in demand in the EMU will result in smaller falls in demand for NMS exports, increases in demand in EMU will result in smaller increases in demand for NMS exports). Thus thanks to ‘vertical II integration’ of this kind, NMS would be less affected by EMU demand shocks than equally II integrated EMU members5. Moreover, it needs to be remembered that II trade also contributes to the convergence of business cycles in a currency area in the presence of aggregate shocks, be they supply or demand shocks. Thus, an increase in aggregate demand in EMU (but not in a particular NMS) will spill over to the NMS through increased demand for its exports, whatever the nature of these. And an asymmetric increase in aggregate costs, constituting a negative aggregate idiosyncratic supply shock (for instance as a result of increased energy prices which affects the NMS more than EMU) will be partly cushioned by the smaller fall in EMU output and therefore demand6. Thus, it is hardly surprising that studies show that business cycles are more correlated between the NMS and Germany than between important EMU members.

WHEN SHOULD THE NEW MEMBER STATES JOIN EMU?

5

Boone and Maurel (1999), using de-trended unemployment, show that between 55% (Poland) and 86% (Hungary) of the NMS cycles are explained by German cycles, whereas only 43% of Spanish and 18% of Italian cycles can be explained in this way. Fidrmuc and Schardax (2000) find that Poland's industrial production is as closely correlated with Germany's as is Austria's, and more so than those of Switzerland or Italy. Hungary and Slovenia's industrial outputs are more closely correlated with Germany's than is that of Italy, although those of the Czech Republic and Slovakia are far less correlated. More recent work by Karhonen and Fidrmuc (2001) improves on previous research by correlating EMU members’ and NMS supply and demand shocks with those of the EMU as a whole7. The research shows that Hungary and Estonia seem to have least to loose by giving up their domestic currency, followed by Poland. The remaining NMS either have very low correlations on both kinds of shock, or reasonable ones on one kind of shock but negative ones of a similar size on the other kind. Finally, Lithuania seems to have quite large negative correlations on both kinds of shock. However, much of the data used in the research comes from quite a few years ago, and since trade integration has proceeded apace, one may suppose that the correlation of NMS’ shocks to EMU shocks (of both kinds) has increased (see Chapter 3). 4. IS THERE A TRADE-OFF BETWEEN REAL AND NOMINAL CONVERGENCE? It is clear that if countries were able to use national monetary policy to mitigate the effects of demand shocks (both negative and positive) and to use exchange rate policy to mitigate the effects of negative supply shocks, the path of GDP would be smoother (which would increase welfare somewhat), and the trend rate of growth might even be somewhat higher (less risk could encourage higher investment). Nevertheless, in general, catch-up is a pretty long-term phenomenon. Convergence between similar regions (US states, Japanese prefectures, European regions) proceeds at between 2% and 4% of the initial gap per annum (Barro and Sala-i-Martin, 1998), which implies a half-life for the productivity gap of between 18 and 36 years. This is not a period over which the availability or otherwise of tools for flexible macro policy is likely to be decisive, even if one were to assume that macro policy smoothes fluctuations rather than augmenting them. But this is exactly where the ‘rub’ lies. Monetary policy in robust young democracies with strong populist movements and limited human capital at the central bank can easily be pro-cyclical rather than anti-cyclical. Exchange rate movements can also be more of a ‘shock generator’ than a shock absorber. A recent study (Habib 2002) finds that emerging market risk premia had a significant impact on exchange rates in the Czech Republic, Hungary and Poland during the period mid-1997 – mid- 2001, and on interest rates in the Czech Republic and Hungary. The volatility of the Czech and Polish exchange rates was also affected by ‘volatility contagion’ coming from the emerging markets and there was also some ‘volatility contagion’ of Czech interest rates. Csermely and Vonnak (2002) show that in

6

JACEK ROSTOWSKI

Hungary the main reason for exchange rate and real interest rate movements has been changes in the risk premium Hungary has had to pay. These changes were due to world market increases in the premium required of emerging markets, an effect which would very largely disappear after EMU accession. Thus, we would argue that a monetary union which eliminated the risk of ‘emerging market contagion’ and made a higher level of capital inflow safe, while at the same time providing a strong, externally guaranteed anti-inflationary framework (which might in turn encourage a higher level of domestic savings), might be expected to increase the long term rate of growth of the NMS, rather than reduce it. Another argument which has been put forward is that NMS need to run fiscal deficits in excess of the limit set by the Maastricht criterion and of the requirements of the Stability and Growth Pact (SGP). One justification for this view is that NMS have poor physical infrastructures, that infrastructural investment can often boost economic growth, and that therefore a higher level of borrowing for public investment needs is justified in the case of NMS (European Economy, 2002, Part V). This is an argument for a certain kind of public expenditure rather than for financing that expenditure through borrowing. Such investment can also be financed by reducing other expenditure or increasing taxation. However, if we were to accept that borrowing was the best way of financing under the special circumstances which NMS face, then rather than require that they choose between EMU membership and an optimal investment/financing mix, it is the Maastricht fiscal deficit criterion which should be altered8. Finally, it needs to be remembered that the requirements for fiscal prudence in the Treaty applies to non-EMU-participating EU member states as well9. 5. SHOULD RAPID CATCH-UP DISQUALIFY NMS FROM EMU MEMBERSHIP? Prominent ESCB officials have argued that not just nominal but also ‘real’ convergence should be required of NMS before they join EMU. Thus FranzChristoph Zeitler, a Bundesbank Board member, has argued that NMS should not be allowed to join EMU before they had achieved GDP per capita levels equivalent to 70% of the EU average (EU hopefuls, 2000, 26 Sept.). In fact, the value or stability of a currency has nothing to do with the income of the people who use it. Were this the case, West European currencies would have been much more stable in the 1970s than in the 1870s, whereas the opposite was true. One argument that has been made is that the Harrod-Balassa-Samuelson (HBS) effect means that NMS are likely to have inflation rates 0-2% above the average of the present EMU. On top of this there may be a demand effect resulting from the demand for non-tradables being more income elastic than the demand for tradables. As incomes rise, demand for non-tradables rises faster, and unless productivity in the sector rises faster than in the whole economy (which is unlikely) non-tradable prices will rise faster than tradable prices. This phenomenon of ‘growth related relative price changes’ also happens in current EMU members, but with faster income growth, it would presumably happen faster in NMS. Adding faster growing

WHEN SHOULD THE NEW MEMBER STATES JOIN EMU?

7

countries with higher HBS and ‘demand effect’ inflation to EMU must increase EMU inflation for any given degree of tightness of monetary policy it is argued. This seems to confront the EMU with an unpleasant dilemma: if the 2% ceiling on inflation is to be maintained for the EMU as a whole, interest rates will have to be higher and inflation in slow growing countries will need to be lower than at present. On the other hand, abandoning the 2% ceiling on inflation and allowing NMS to have inflation rates well above the average of present EMU members (without reducing inflation in the current member states) would, it is claimed, undermine price stability in the whole zone. However, we have to take a number of other considerations into account. First, both the HBS effect proper and the demand effect, are as likely to occur in any rapidly growing economy as in a NMS. Indeed, these effects have appeared quite strongly in Ireland, and to a lesser extent in Spain. They are not particularly the effect of catch-up, but simply of faster growth in the country concerned than the EMU on average. Thus, in order to prevent these effects from raising the average inflation rate in the EMU, members should not only refuse to accept NMS, but they should also prevent current members from unilaterally undertaking structural reforms which would accelerate their growth. Indeed, such a moratorium on reform would be particularly important since the NMS represent only about 6.5% of EMU GDP. Thus, structural reforms that increased the growth rate by half in France and Italy (which together account for about 40% of EMU GDP) would be likely to have a much larger effect on EMU inflation than the accession of the NMS. Therefore, unless the EMU wishes to remain a low growth zone, there can be no argument for excluding NMS on such a basis. Indeed, it may be in recognition of this fact, that the Maastricht inflation criterion only needs to be fulfilled for one year. Although the reduction in inflation does need to be sustainable to satisfy the inflation criterion, it is hard to imagine that expectations of ‘growth generated relative price changes’ of the kind we are discussing after the conclusion of the ‘reference period’, could be held to violate this condition10. Second, the NMS are so small economically that their accession to EMU would not in fact require any change in the ECB’s inflation target, even if their ‘growth inflation’ were relatively high. Thus, if NMS inflation were an improbably high 3 percentage points above the average of current EMU members, this would mean an increase in enlarged EMU inflation of 0.2%, on the assumption that there were no offsetting effects. If France and Italy increased their long-term growth rate by half, the effect could be twice as large (if we assume that the inflation effect of a growth acceleration would be about half of the acceleration). Of course, even if the EMUwide inflation target were adjusted upwards to take ‘growth inflation’ in parts of the zone into account, there would not be any need to adjust the ECB’s monetary policy, and therefore there should not be any effect on the inflation rate in slow growing countries. Third, it is not in fact clear that either the HBS effect or the demand effect on non-tradables prices in the faster growing countries, would in fact increase the inflation rate in EMU as a whole, once we take the impact of the higher growth on the nominal exchange rate of the euro into account.

8

JACEK ROSTOWSKI

To see this, let us imagine that all EMU countries increase their growth rates by half. Would this lead to higher inflation within EMU? Assuming a stability- oriented monetary policy there are four ways of thinking about what would happen: x To the extent that the higher growth was mainly generated by productivity increases in the traded sector (leading to an HBS effect), it would lead to a nominal appreciation of the euro against other currencies (we assume that the growth acceleration would be limited to the EMU). Domestic EMU prices of tradables would fall (or fall more rapidly than they are in any case doing before the acceleration), compensating for the increase in the prices of non-tradables and overall inflation would remain roughly constant11. x To the extent to which the productivity growth occurred in non-tradables, prices in this sector would not rise in the first place12. x To the extent to which we have a shift in the structure of demand from tradables to non-tradables, the euro zone’s balance of payments will improve ceteris paribus, leading to nominal appreciation of the euro, reducing the euro price of tradables and thus compensating for the increase in non-tradables prices13. The nominal appreciation also increases the marginal costs of producing tradables in the EMU relative to their world price (in dollars) and leads to resources being transferred inside the euro area to non-tradables production. x To the extent to which higher growth is balanced across sectors and there is no ‘demand effect’, prices of non-tradables would not rise. However, higher growth would lead to pressure for higher real interest rates, which would also lead to nominal appreciation of the euro, which would cause the price of tradables to decline, and overall inflation actually to fall. We can now consider what happens when the increase in growth is not EMUwide. In the case when it is limited only to France and Italy, we would have weaker HBS or ‘demand effect’ impacts on EMU-wide non-tradables prices, more or less compensated by an equivalently weaker nominal appreciation of the euro. In the last case above, of ‘fully balanced’ growth acceleration (on both the demand and supply sides in both sectors) in France and Italy, the real interest rate effect on tradables prices would also be weaker than in the case of an equally large EMU-wide growth acceleration. In all of these scenarios the impact of such a ‘geographically unbalanced’ growth acceleration on slower growing countries within EMU could nevertheless be quite unpleasant. Nominal appreciation of the euro would make some part of their tradables output non-competitive at previously expected wages. We can see this if we initially assume the existence a single homogeneous tradable good, for which world demand and world supply are perfectly elastic. This good is produced in different currency areas of the world according to production functions which differ between areas in the short term, and which generate increasing marginal costs in production. Area output is determined at the level at which the world price PT = MCT (domestic marginal cost). Resources (at least two factors) used in the production of the tradable good are also used in the production of each area’s nontradable good. An increase in productivity in the production of the tradable good in

WHEN SHOULD THE NEW MEMBER STATES JOIN EMU?

9

an area leads to an increase in area income and therefore to a change in the structure of demand. Assuming that this causes a greater increase in demand for the nontradable good in the area than for the tradable, the price of the non-tradable good will rise. This implies a real appreciation of the area’s currency, and in a noninflationary environment, a nominal appreciation of the currency as well. Resources are then transferred from production of the tradable to the non-tradable good, which will additionally benefit the suppliers of the factors of production which are used more intensively in the production of the non-tradable good (by the Rybczynski Theorem). ‘Regions’ within the currency area (in the EMU case these could be countries such as Germany) in which productivity in the production of the tradable good does not increase, will find that with nominal appreciation, domestic PT will fall. If domestic factor prices are flexible in such a region, this may cause a fall in nominal returns to factors of production which are relatively immobile out of the region (labor). If domestic factor prices are downwardly sticky, it will cause a fall in the output of the tradable good in the region, and a reduction in employment. Thus, in order to avoid an increase in unemployment, expected wage increases might have to be foregone, or in the case of a relatively strong effect, real (or even nominal) wages might have to fall. This latter case is made the more likely by the low level of the ECB’s inflation target. Of course, on average even the inhabitants of slow growing EMU-members will benefit from faster growth anywhere in EMU (or indeed in the world). But these gains are likely to be unevenly distributed, with consumers of tradable goods benefiting more than their producers. Thus, if we assume two tradable goods are produced in the world, the currency area and the region we have been considering, and that productivity increases in the production of only one of the two tradable goods in one of the regions of our area (but not in the other region or the rest of the world) then we get the following effects: (1) the area’s currency still appreciates; (2) in the region in which productivity has increased, resources are transferred from the production of the both tradable goods to the non-tradable good; (3) in the region in which productivity has not increased, there are two rounds of effects. The first round involves, as before, a reduction in the profit maximizing level of output of both tradable goods at given factor prices. This requires either a fall in output and an increase in unemployment, or a fall in regional factor prices (relative to other regions and in nominal terms in a non-inflationary environment). The second round is a consequence of the overall currency area growth which has resulted from the increased productivity in the production of one of the tradable goods in the other region of the area. This increases area demand for the other tradable good (in whose production productivity has not increased), which increases demand for the factors used intensively in the production of this second tradable good. The non-growing region should benefit from this, as it can increase exports of the second tradable good to the growing region. Which of these two (first or second round) effects will predominate depends on the size of the slow growing region relative to the area and on the various elasticities involved. If factor prices are highly flexible downwards, then the region will not loose from the first round effect and will gain (albeit maybe only slightly) from the second round effect.

10

JACEK ROSTOWSKI

In the case of NMS accession and subsequent catch up, the direction of the effects on EMU slow growers would be the same as in the two cases examined above, but their magnitude is likely to be far smaller14. Thus, to summarize, the problem with NMS accession to EMU is not that it will cause higher inflation in the whole euro zone, but rather that it may cause higher unemployment in slow growing countries such as Germany if they fail to increase the flexibility of their labor markets (see Chapter 10). 6. EMU INSTITUTIONAL ISSUES One reason for the Maastricht criteria for EMU accession is that EMU membership gives a country voting rights in the ECB and the Eurogroup of the Ecofin, and existing members would not want countries that have not demonstrated a commitment to low inflation to have such a vote. This is especially important in the case of the ECB. Although the ECB is supposed to make policy for the EMU as a whole, and therefore small economies should have a small weight in its considerations, each member of the interest rate setting ECB Governing Council (GC) has one vote at present. Apart from the six Executive Board (EB) members, the governors of all the national central banks (NCBs) are members of the GC (giving a total of 18 members in the GC). This is a concern given that there are at present 12 potential new members from central, eastern and southern Europe. The Nice summit of December 2000 asked the ECB to make proposals for the restructuring of the Governing Council, and the ECB has now done this. The current member states of the EMU may be concerned that most of the potential new members have recent histories of far higher inflation than Western Europe. The fear could be that they would not have the same commitment to price stability as existing members. We have called this the ‘price-stability culture problem’ (Bratkowski and Rostowski, 2002). This problem is partly offset by the fact that central bankers are not as a breed prone to be inflation lovers. Second, all European System of Central Banks (ESCB) member banks now have to be highly independent for countries to qualify for EU, let alone EMU, accession. Indeed, the opposite fear has also been expressed: central bankers of NMS might vote for shortterm interest rates which kept inflation under control in their own countries, but were too high for core countries such as Germany. It has also been claimed that a GC of 28 members would be far too large for serious discussion of the merits of monetary policy issues, which are best decided in a small group. It is said that in the ECB the tradition has been not to vote but to reach consensus and, it has been argued, reaching consensus in a group of twenty odd could be time consuming. Neither of these arguments seems to have any real merit. It is impossible to believe that consensus is at present arrived at in the 18 member Council through general discussion. What seems far more likely is that there is a smaller group with a clear view, and the rest go along with that. Compared to this, voting at least has the advantage of transparency. Moreover, the practical irrelevance of this argument has been demonstrated by the ECB’s own proposal for

WHEN SHOULD THE NEW MEMBER STATES JOIN EMU?

11

the reform, which foresees a GC of 21 voting members, and that all ‘National Central Bank Governors will continue to participate in the discussions of the Governing Council and attend the meetings…’ - which implies up to 33 participants (ECB, 2002). Third, there is the serious matter of the unfairness of the present arrangement, which with its ‘one governor one vote’ principle, favors small countries enormously, even if all the members of the executive board were drawn from large countries (which has not been the case). The admission of the NMS merely exacerbates and highlights the problem. Nevertheless, this problem has now been largely solved by the ECB’s proposal for the reform of the GC. This envisions 6 board members and 15 rotating country representatives. The five largest countries in terms of their shares of GDP will have four votes rotating among them15. The remaining countries will be divided into two groups as follows: the next largest one half of all the EMUparticipating countries by GDP will have eight votes to rotate amongst themselves, while all the remaining countries will have three votes to rotate among themselves16. It should also be pointed out that whereas it is not in fact very dangerous, from the point of view of maintaining the commitment to price stability, to admit NMS central bankers to the ECB GC (given central bankers usual preferences), the same cannot necessarily be said as regards the effects on sound public finance of admitting NMS finance ministers to a powerful ‘EMU-Ecofin’. The risk of doing so may be perceived as even greater if the SGP is seriously watered down (see Chapter 9). 7. CONCLUSION Thus, the balance of the argument appears to lie definitely on the side of EMU accession when we look at it from the perspective of the NMS interests. However, from the perspective of incumbent EMU members’ interests, although we can state that many of the arguments put forward against early EMU accession have been overblown, we cannot avoid the conclusion that the arguments are more finely balanced. This is especially the case when we analyze the situation of slow growing incumbents, who are also likely to be members of any ‘core’ grouping, and that of a Commission that is eager to expand its powers into fiscal policy and fiscal policy coordination. Given the opposing interests between the two sides, and the fact that it is impossible to forbid EU members from joining EMU if they fulfill the Maastricht criteria, we can expect to see an intensified effort by the Commission and certain incumbent Member States to dissuade NMS from joining EMU. There may even be an attempt to return to the idea of tightening the accession criteria, although this would be hard to square with the principle of ‘equality between present and future Member States’17.

12

JACEK ROSTOWSKI

NOTES 1

Estonia has satisfied the Maastricht Treaty’s exchange rate criterion ever since the ratification of the Treaty and the interest rate criterion for many years. These are the two criteria which need to be satisfied for two years. The exchange rate criterion requires that the value of the applicant country’s currency remain within the range required by the ERM2 relative to the central parity (+/- 15%) for a period of two years, without a devaluation of the central parity. The interest rate criterion requires that interest rates on long term (ten year) government bonds not exceed the average of the three countries with the lowest inflation by more than 2 percentage points. The remaining criteria are: (1) that inflation should not exceed the average of that in the three best performing members of the EMU by more than 1.5% for one year; (2) that the fiscal deficit should not exceed 3% of GDP, and (3) that gross public debt should not exceed 60% of GDP. 2 The US experience in the 1990s is instructive in this context. 3 Most of the NMS are small countries in terms of population and very small in terms of GDP, and small countries tend to have less diversified foreign trade. With the exception of Finland and Greece it is the larger EMU members that have lower levels of international trade (and therefore of EMU trade/GDP). They may, therefore offset their lower levels of trade with higher levels of diversification. However, it is not clear whether in NMS the extent of diversification depends on GDP or population. Also, small countries may be less affected by a given asymmetric shock than large ones. Usually OCA theory assumes that all goods produced by a given sector are identical, irrespective of which country they were produced in. However, if there are specificities to the goods produced by different countries, each country will face a somewhat downward sloping demand curve for its exports. Such demand curves are likely to be more elastic in the case of smaller countries, since they supply a smaller share of the market for their own produce and that of close competitors. A negative asymmetric supply shock to one sector can be compensated by an expansion in a country’s other exports, and this is likely to be easier in the case a smaller country. Further empirical study of this question is needed. This problem is likely to have less relevance for the second measure of trade with EMU/total trade, where it is the largest NMS which score highest, and they do so against almost all EMU members. 4 NMS’ scores on both measures would be somewhat higher if we were considering their exposure to shocks after all of them had joined EMU, since NMS also trade with each other. 5 On the reasonable assumption that any monetary policy response to such a shock for the EMU as a whole would not fully offset the effects of the shock. 6 This is not to deny that the effects of such a shock might be even more muted if the NMS currency could depreciate in such a situation. But we are not talking here of the choice between accession or not for the NMS, but rather about whether they are less suited to accession than current EMU members. It is in this context that it is important to remember the role of II trade in muting the effects of shocks even in a currency union. 7 The period covered depends on the availability of statistics, but for EMU countries it is for 1991 Q3 – 2000 Q3 or Q4. For NMS it ends in 2000 Q4 (except for Slovakia, which ends in 2000 Q3) and begins as early as 1992 Q1 (Romania) or as late as 1995 Q2 (Estonia, Hungary, Latvia and Lithuania). 8 A variant of this argument is that the EU itself requires various expenditures (for instance on environmental protection). These can be thought of as the ‘admission price’ to the EU, which is worth paying because other aspects of accession will boost growth in such a way as to more than compensate for it. If the optimal method of financing these expenditures were borrowing, then rather than excluding the NMS from EMU or forcing them to finance the admission price in a sub-optimal way, it would again be better to adjust the fiscal deficit criterion appropriately. 9 All EU members are obliged to avoid excessive deficits (defined as those in excess of 3% of GDP) under the Treaty, whether they are EMU members or not (the sole exception is once again the UK, which is only obliged to ‘attempt to avoid’ an excessive deficit). Only EMU members can be fined for failing, but the legal requirement nevertheless remains. Furthermore, any member state which has an excessive deficit should be unable to obtain financing for any new projects under the EU Cohesion Fund, something which is of great importance for NMS (Ecofin 1999, Art 2, Para 4 and Art 6, Para 1). Finally, all EU

WHEN SHOULD THE NEW MEMBER STATES JOIN EMU?

13

members are obliged to achieve ‘a fiscal position close to balance or in surplus’ in accordance with the SGP. Neither EMU nor non-EMU members can be fined for failing to fulfill this requirement either, although again both will loose access to the Cohesion Fund if they are found by the Council to be in breach of the Pact (Ecofin 1999, Art 1, Para 1). 10 The ‘reference period’ is that for which the conformity of a country with the criteria is assessed. What could be targeted by the sustainability requirement, would be various tricks such as ‘delaying’ administered price or negotiated wage increases into ‘post-examination’ years, or the fulfillment of the criterion thanks to an unsustainable nominal appreciation. 11 As long as the domestic price elasticity of demand for both categories of goods (tradables and nontradables) were the same. 12 As long as the ‘structure of demand’ effect on the demand for non-tradables, resulting from the higher income the accelerated growth brings, is not larger than the growth in productivity in non-tradables. 13 The shift is in the structure of demand, so euro zone demand for tradables is likely to grow in absolute terms. This is because the shift itself results from overall growth in the economy. 14 Throughout the analysis we assume that the EMU does not face a downward sloping demand curve for its products. If it does, then productivity growth in the tradables sector will lead to a nominal depreciation of the euro rather than an appreciation, whether it is balanced by equal productivity growth in nontradables or not. Nominal depreciation will put upward pressure on EMU inflation, which would need to be offset by higher nominal interest rates. If the tradable goods produced by all EMU countries were homogeneous, producers in countries with slow productivity growth would be hit by competition from their neighbors and by higher real interest rates. To the extent to which the tradables produced by different countries differed, producers in slower growing EMU countries would face only the second problem of higher real exchange rates. However, only areas with largely undiversified exports (such as raw materials producers) are likely to face demand curves which are downward sloping in the medium to long term (the time horizon which concerns us, as we are considering issues of differential growth rates). 15 The actual criterion is a composite indicator of each country’s supposed ‘representativeness’ of the EMU economy as a whole. This indicator is composed in 5/6 of GDP and in 1/6 of the total assets of the aggregated balance sheet of monetary financial institutions (TABS-MFI). It is unlikely that TABS-MFI will significantly affect the allocation of Member States to the three groups. 16 The third group will only come into existence when the number of states participating in EMU reaches 22. 17 This principle is claimed to be the basis for the Commission’s opposition to unilateral euroization (European Commission, 2000).

REFERENCES Barro, R. & Sala-i-Martin, X. (1998). Economic growth. MIT Press. Cambridge Mass., pp. 583. Boone, L. & Maurel, M. (1999). An optimal currency area perspective of the EU enlargement to the CEECs. Discussion Paper, 2119, Centre for Economic Policy Research, London. Bratkowski, A. & Rostowski, J. (2000). Unilateral adoption of the euro by EU applicant countries: the macroeconomic aspects. From Ten Years of Transition, ed. Orlowski L. Edward Elgar. Bratkowski, A. & Rostowski R. (2002). The EU attitude to unilateral euroization: misunderstandings, real concerns and sub-optimal admission criteria. Economics of Transition, 10 (2), 445-468. Calvo, G. (1999). On Dollarization. http://www.bsos.umd.edu/econ/ciecpn5.pdf Commission warns applicants on EMU (2002, October 10). Financial Times. Csermely, A. & Vonnak B. (2002, October 3). The role of the exchange rate in the transmission mechanism in Hungary. Paper presented at the research meeting on ‘Monetary policy transmission in the Euro area and the accession countries’. National Bank of Hungary, Budapest. ECB (2002, December 20). Governing council prepares for enlargement. Press Release, European Central Bank, Frankfurt. Ecofin (1999, June 26). Council regulation (EC), 1264/1999. Official Journal of the European Communities, L161/57. EU hopefuls likely to suffer if Danes vote ‘no’ (2000, September 26). Financial Times. European Commission (2000, August 22). Exchange rate regimes for applicant countries. Note for the economic and financial committee, Brussels, mimeo, ECFIN/521/2000-EN.

14

JACEK ROSTOWSKI

European Economy (2002, June 3). Public finances in EMU - 2002. European Economy, European Commission, Directorate-General for Economic and Financial Affairs, Brussels, from http://europa.eu.int/comm/economy_finance/publications/european_economy/2002/ee302en.pdf Fidrmuc, J. & Schardax F. (2000). More ‘Pre-Ins’ ante portas? Euro area enlargement, optimum currency area, and nominal convergence, Transition, 2, 28-47. National Bank of Austria, Vienna. Habib, M. (2002). Financial contagion, interest rates and the role of the exchange rate as a shock absorber in Central and Eastern Europe. Discussion Paper, 7. Bank of Finland Institute for Economies in Transition (BOFIT), Helsinki. Karhonen, I. & Fidrmuc J. (2001). Similarity of demand and supply shocks between the Euro area and the accession countries. Focus on Transition, 2, 26-42, National Bank of Austria, Vienna. Rostowski, J. (2000). The approach to EU and EMU membership: the implications for macroeconomic policy in the applicant countries. In The Eastern Enlargement of the European Union eds. Dąbrowski, M. & Rostowski, J. Kluwer Academic Publishers. Sulling, A. (2002). Should Estonia euroize. Economics of Transition, 10 (2), 469-480.

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

CHAPTER 2

THE EXCHANGE RATE: SHOCK GENERATOR OR SHOCK ABSORBER?

1. INTRODUCTION The aim of this study is to assess the impact of exchange rate regimes on inflation and per capita growth in a wide group of countries. Given a wide range of factors that influence the choice and effectiveness of the exchange rate regime, we turn to empirical analysis to see which regimes have been associated with lower output volatility, attempting to answer the question whether flexible exchange regimes insulate economies from shocks or generate disturbances oF their own. The study draws on Ghosh, Gulde and Wolf (2002) (GGW), who conduct a similar exercise employing the exchange rate classification based on the IMF Annual Report on Exchange Rate Arrangements and Exchange Restrictions. GGW find that inflation is lower under pegged regimes. This result reflects both a greater discipline imposed on a central bank – reflected in lower monetary growth – and a higher credibility of the system, reflected in a lower velocity of money. Flexible exchange rate arrangements are, on the other hand, associated with a lower variance of output. The lower inflation delivered by the pegged regimes comes therefore at the cost of higher real volatility. There seems to be no strong link between the per capita output growth and the exchange rate regime. The official classification, however, takes no notice of the de facto exchange rate behavior, treating equally exchange rate pegs subject to frequent and infrequent adjustments. GGW approach this problem by constructing a ‘consensus’ classification, dropping cases where the actual exchange rate behavior is markedly different from the official classification. Our paper applies the ‘natural classification’ of Reinhart and Rogoff (2002) (RR), which is based on the behavior of exchange rates – either official or parallel – that are predominant in the economy. We conduct an extensive comparison of results across various classifications and samples. The actual behavior of the exchange rate may reflect shocks affecting the economy, rather than the authorities’ attempts to affect exchange rate movements. The ‘natural’ classification does not distinguish between these two cases, creating problems for the interpretation of the relationship between macroeconomic variables and exchange rate arrangements. In the absence of shocks, it is more likely that 15 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 15-40. © 2006 Springer. Printed in the Netherlands.

16

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

economic performance will be above average, and that exchange rate behavior will be classified as a variant of a regime with limited flexibility. The relationship between the exchange rate regime and economic performance may therefore be spurious, reflecting the impact of shocks that are common to the economy and to the exchange rate classification. We address this problem, which may be particularly acute when using the ‘natural’ classification, by checking for the robustness of our analysis under alternative specifications of the inflation equation. The chapter is organized as follows. Section 2.2 describes the theoretical considerations pertaining to the choice of exchange rate regime. Section 2.3 explains the RR classification adopted in our study. Section 2.4 discusses the data. Section 2.5 describes the methodology, providing a further motivation for the use of the ‘natural’ classification and for handling regime endogeneity. Section 2.6 reports and discusses empirical results, and Section 2.7 concludes. 2. THE THEORY OF EXCHANGE RATE REGIME CHOICE The literature on the choice between fixed and flexible exchange rates can be grouped into three broad categories. The first focuses on the insulating properties of regimes. The second examines the impact of different exchange rate regimes on economic integration. Two main issues here are whether the fixed exchange rates reduce uncertainty and transaction costs thereby leading to greater economic integration and what are the conditions under which it is preferable for a group of countries to forgo a domestic monetary policy and form a currency union. The third strand focuses on the credibility aspect of the monetary regime. 2.1 Transmission of shocks to the real economy The literature on how various regimes would operate under conditions of high capital mobility derives from Fleming (1962) and Mundell (1963). These papers point to different implications of fixed and floating exchange rates for the conduct of stabilization policy. If an economy faces mainly nominal shocks then a fixed exchange rate regime looks more attractive. If the shocks are real e.g. like productivity shocks or changes in the terms of trade, the economy needs to react to changes in relative prices. A flexible exchange rate regime allows for a quick change in relative prices, which reduces the impact of the shock on output and employment. The combination of shocks under which fixed exchange rates would be preferable to floating depends on capital mobility. If capital is relatively immobile, fixed exchange rates provide better insulation of output against shocks to aggregate demand, while under high capital mobility flexible exchange rate regime is preferable. This is due to asymmetric impact of trade and capital flows on the balance of payments. If capital mobility is low, under fixed exchange rates a positive shock to aggregate demand leads to higher imports and loss of reserves through a trade deficit. If the loss of reserves is not sterilized the money supply contracts partly offsetting the original shock. Under a floating regime the trade deficit leads to

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

17

depreciation of the exchange rate, which leads to higher exports magnifying the initial shock. When capital is highly mobile, the balance of payments effects dominate. Under fixed exchange rates, the positive demand shock raises interest rates, which results in capital inflow that more than compensates for the loss of reserves due to the trade deficit. Therefore money supply is higher magnifying the initial shock. Under a floating exchange rate the capital inflow appreciates the exchange rate, this reduces exports and partly offsets the initial shock. 2.2 Economic integration Adopting a “hard-pegged” exchange rate implies that the nominal exchange rate can no longer serve as an instrument of adjustment. Therefore the case for adoption of a common currency is stronger if countries are subject to similar shocks. This is the main finding of the optimum currency area literature originating from Mundell (1961). The loss of the adjustment mechanism of the exchange rate is less significant if other adjustment mechanisms are available (such as wage and price flexibility, factor mobility, and fiscal transfer systems). The gains from a hard peg include increased trade and investment flows. 2.3. Credibility The main advantage of a floating exchange rate is that it provides the ability to employ monetary policy to cope with shocks to the domestic economy. However it can at the same time be criticized for allowing too much discretion in monetary policy, and for not providing a sufficiently strong nominal anchor. In a closed economy the central bank can pre-commit to low inflation by relying on the repeated game nature of the interaction with wage setters or on the appointment of a hawkish central banker (in an independent central bank). In an open economy, an alternative pre-commitment strategy is the pegging of the nominal exchange rate to a low-inflation country. The peg then imposes an additional constraint on the central bank’s ability to create inflation surprises. This constraint is credible as long as the perceived costs of abandoning the peg are greater then the benefits of generating surprise inflation. Adoption of the hard peg can thus increase credibility and make it easier for the central bank to achieve and maintain low inflation. 2.4. Fiscal constraints, financial fragility and other considerations Under a fixed exchange rate regime with high capital mobility, fiscal policy is the only tool of macroeconomic stabilization. Therefore the ability to employ fiscal policy as an adjustment mechanism is one of the factors that one needs to consider in choosing the exchange rate regime. High fiscal deficits or debt ratios can undermine the credibility of the government, as investors might expect that the government will seek to monetize the deficit abandoning the peg in the process. The

18

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

fiscal theory of price determination applied to the exchange rate regime (Canzoneri, Cumby and Diba, 1998) suggests that the fixed exchange rate will be sustainable only if fiscal policy is sufficiently flexible to respect the government’s present value budget constraint at a price level consistent with the exchange rate peg. On the other hand, some models suggest that fixed exchange rates create an incentive for a government with a short time horizon to run larger deficits and deliver short term growth with inflationary costs imposed on future governments (Tornell and Velasco, 2000). Another issue to be considered is the ability of the central bank to act as a lender of last resort. Hard exchange rate pegs do not allow the central bank to act as a money-printing lender of last resort. However, in the case of emerging-market economies this does not seem to be a genuine concern. While in developed countries the monetary authority can issue liquidity to bail out the banking system, this extra liquidity is expected to be absorbed in the near future by open market operations without inflationary consequences. In emerging-market economies, central bank lending to the banking system in a wake of a financial crisis and a sudden stop in capital inflows is likely to unleash fears of an inflationary explosion and to lead to a sharp depreciation of the exchange rate. If a large proportion of private debt is denominated in foreign currency, this will lead to even more financial instability. Further, even if a country is better off with a floating exchange rate, the shift from a fixed regime might have serious economic consequences. The move from peg to floating regime in the midst of a crisis is likely to exacerbate the crisis. The initial devaluation, which raises the value of foreign-denominated debt, can cause widespread destruction of private balance sheets, which can lead to a downward spiral. In addition, restoring the national currency might also lead to a major overhaul of the domestic financial sector (Caballero and Krishnamnurthy, 2002; Jeanne, 2002). 3. CLASSIFICATION OF EXCHANGE RATE REGIMES It is widely recognized that exchange rate regime classifications based on declarations of central banks do not correctly depict reality. Officially pegged exchange rates are often subject to frequent adjustments. On the other hand, Calvo and Reinhart (2000) and Hausmann, Panizza and Stein (2001) claim that countries with de jure flexible exchange rates regimes often do not allow their exchange rates to move freely (‘fear of floating’). Factors such as a lack of credibility, combined with a high exchange rate effect on inflation (pass-through) and the potentially devastating impact of large exchange rate changes on the banking sector, may prevent countries from pursuing an independent monetary policy. There is evidence that in the developing countries in which authorities defend the exchange rate without a formal commitment, interest rates are more sensitive to changes in the US interest rate than in countries with officially fixed exchange rates. Such “floaters’” risk premiums may thus be more sensitive to the US interest rates, requiring stronger interest rate adjustments to reduce exchange rate volatility, and further limiting the scope for independent monetary policy.

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

19

In addition to the vast difference between de jure and de facto exchange rate behavior, relying on the behavior of officially reported exchange rates may be also misleading. Most countries after World War II relied on capital controls and/or multiple exchange rate systems at some stage, and it is not possible to assess the underlying monetary policy of a country or the ability of an economy to adjust imbalances, without looking at the market-determined exchange rate. RR identify the periods when dual or multiple exchange rates were in place or when parallel markets were active. They use monthly data on parallel/dual exchange rates to check for consistency between the de facto and de jure exchange rate regimes and construct a ‘natural’ classification, based on the actual behavior of the predominant – either official or parallel – exchange rate. They find that under the Bretton-Woods system many countries had a de facto floating exchange rate. In about 45% of arrangements officially classified as pegs, the actual regime was in fact a managed or freely floating arrangement with limited flexibility. On the other hand, RR classify 53% of the regimes officially classified as managed floating as in reality being pegs, crawls or narrow bands to an anchor currency. Under the ‘natural’ classification, the most popular exchange rate regime during 1970-2001 was the peg (33% of observations based on 153 countries), followed by the crawling peg or narrow crawling band (26% of observations). Finally, RR stress the importance of distinguishing the countries with inflation over 40%, and classify these countries as having a freely falling regime. Relationships between macroeconomic variables in these high-inflation countries are very different from those in more stable economies. It is therefore important to check for the robustness of any estimated economic relationships when the highinflation cases are excluded. RR do not provide an econometric analysis of the impact of the re-classified exchange rate regimes on inflation and growth. However, the reported average per capita growth rate for the freely floating regimes under the standard classification is 0.5%. Once the freely falling regimes are separated from the free-floating regimes, the growth rate of countries with freely floating exchange rates increases to 2.3%. 4. DATA The RR classification covers 144 countries over the 1940-2001 period. Data on inflation, GDP growth and a range of additional controls, such as broad money growth, terms of trade, dollar value of exports and imports, investment share in GDP, are reported in GGW and originate from the World Economic Outlook database. The official classification of exchange rate regimes is also taken from the data reported in GGW. Data availability differs across countries: neither the RR nor GGW classifications cover all of the countries across the whole period. For the sake of comparison, in the case of both inflation and growth we use the same sub-sample of observations as in GGW, excluding observations not available in the RR classification. As pointed out in RR, comparisons of pegged and floating regimes that do not separate the freely falling cases are meaningless. Since we focus on

20

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

countries with low and moderate inflation levels, hyper-inflationary cases – i.e. observations that belong to the ‘freely falling’ RR group – are excluded in most of the samples. In addition, we exclude from the same samples the RR category called ‘dual market in which parallel market data is missing’. 5. INFLATION AND GDP GROWTH ACCORDING TO THE RR CLASSIFICATION Table 2.1 reports the average inflation rate in countries grouped according to the RR classification. The best inflation performance, with average annual inflation of 4.9%, is recorded in countries with no separate legal tender. This is followed by the ‘de facto peg’ regime, with an average of 5.3%, and the ‘moving narrow band’ of 6.5%. In the ‘freely falling’ regime, which by definition records the highest inflation, the average annual CPI growth is 302.5%. The second worst group is the pre-announced crawling peg with average inflation of 55%. The best growth performance (see Table 2.2) is recorded in the residual category (‘dual market in which parallel market data is missing’). High GDP growth is also recorded in the narrow moving band, the de facto crawling band and the preannounced peg (all around 5%). A similar growth performance is recorded on average in countries with no separate legal tender. Overall, it seems that this category of countries seems to enjoy the lowest level of inflation and a growth rate comparable to the best performance in other regimes, although the inflation performance of the pegged regimes seems to deteriorate towards the end of the sample. The pegged exchange rate regimes are more heavily concentrated in the lowinflation 1960s, while the flexible regime observations are predominant in the higher-inflation 1970s and 1980s. The shift from low inflation under pegged regimes to high inflation under floating might be explained by the choice of exchange rate regime, but it might equally well be attributed to negative macroeconomic shocks in the later period. Table 2.3 summarizes information provided above. Categories are combined into six groups, of which the first three are aggregates of the fine classification (each consisting of 4 consecutive subgroups), and the last three are the same as in the fine classification. On balance, the pegged regimes do best in terms of combined performance of inflation and GDP growth. 6. METHODOLOGY Our aim is to test for a relationship between exchange rate arrangements (defined by the ‘official’ and ‘natural’ classifications) on the one hand, and inflation, per capita growth, and the volatility of output on the other.

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

21

6.1 Inflation The commonly held view is that pegged exchange rates can be an important antiinflationary tool. This is because a pegged exchange rate provides a visible commitment and raises the costs of excessive monetary growth. Also, if the peg is credible, the increase in money demand following given monetary expansion is likely to be bigger. Our basic specification – identical to GGW – models inflation by the inverted money demand equation, adding other determinants:

S

E 0  E ExrR ExrR  E m 'm  E y 'y  E Open Open 

E CBTTurn CBTTurn  E ToT 'ToT  E GovGDP GovGDP  H

(2.1)

where ExrR stands for exchange rate regimes dummies, 'm is broad money, 'y is real output, Open is trade openness, CBTurn is turnover rate of the central bank governors, 'ToT stands for changes in the terms of trade, and GovGDP is the fiscal balance as a % of GDP. The trade openness enters the inflation equation since it raises the costs of monetary expansion, as argued by Romer (1993). The turnover rate of central bank governors is a measure of central bank independence proposed by Cukierman (1992). Higher central bank independence helps solve the time inconsistency problem, leading to lower money growth and lower inflation. International terms of trade are an exogenous source of inflation (Fischer, 1993). The government deficit can affect inflation both through direct monetary financing of the deficit and through the pressure from increased aggregate demand. The basic specification is augmented with aggregated and disaggregated regime dummies, time dummies to capture unobservable shocks common to all countries and country dummies. We use this formulation to compare results with the RR and GGW classifications in the two samples: one including all categories and the second excluding the ‘freely falling’ and the residual category as defined in the RR classification. All subsequent regressions use the second sample excluding the ‘freely falling’ and the residual category. The above formulation of the inflation equation implies that explanatory variables other than money and GDP growth rates explain changes in the velocity of money. The indirect influence of the additional variables on inflation is purged from regression coefficients by including money as a separate explanatory variable. In our second specification the money growth rate is excluded. The models, both including and excluding money growth, include potentially endogenous variables, leading to considerable methodological difficulties. Estimation of the model depends on assumptions about the structure of the error term, correlation between errors and explanatory variables, as well as about the equation’s dynamics. Consider the following general form of the panel data model:

y it

Exit  K i  X it ;

i 1,2,..., N ; t

2,3,..., T ;

(2.2)

22

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

There are several factors potentially affecting inflation performance, which are country-specific and either unobservable (e.g. society’s aversion to inflation) or difficult to measure for a diverse group of countries (e.g. degree of wage indexation). It is therefore reasonable to assume a non-zero variance for unobservable components Și. Parameters of the model can be consistently estimated by OLS if E ( x' it X it ) 0 and E ( x ' it K i ) 0 for t 1,2,..., T . But both of these conditions are likely to be violated: money growth, output growth and government balance can be determined simultaneously with inflation; and unobservable country characteristics can be correlated with the choice of exchange rate regime and the behavior of other variables. The fixed-effects transformation of the model averages equation 2.2 over time:

yi

E xi  K i  X i

(2.3)

and subtracts the cross-section equation 2.3 from the original equation:

y it  y i

E ( xit  xi )  X it  X i ; i 1,2,..., N ; t

2,3,..., T ; (2.4)

This transformation removes individual specific effects Și from the model, and estimation of equation 2.4 by OLS eliminates a potential bias from the correlation of unobservable characteristics with explanatory variables. The fixed-effects estimation, however, does not eliminate correlations between explanatory variables and idiosyncratic shocks Xit. The instrumental variables (IV) method allows for a consistent estimation of the model, but a special structure of the transformed error term in equation 2.4 requires instruments zit to be strictly exogenous conditional on fixed-effects, i.e.:

E (X it z i1 ! z iT ,K i )

0

(2.5)

This assumption implies that shocks Xit are uncorrelated with past, current, and future instruments. The assumption is stronger than the earlier assumption of no contemporaneous correlation between errors and explanatory variables sufficient for consistency of the pooled OLS estimation. Any correlation of Xit with past or future instruments renders them invalid, since the transformed equation 2.4 contains timeaveraged errors X i . This, in particular, excludes lagged values of endogenous explanatory variables from the list of valid instruments. Similarly, when an endogenous explanatory variable is replaced in the model by its lagged values, the correlation between variable’s future values and shocks renders the estimates inconsistent. Estimation of a model with lagged dependent variables encounters the same problem since a lagged dependent variable is correlated with time-averaged shocks in equation 2.4.

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

23

An alternative approach to eliminating unobserved effects is first differencing of equation 2.2:

'y it

E'xit  'X it ;

i 1,2,..., N ; t

2,3,..., T ; (2.6)

The new transformed equation is more amenable to IV using appropriate lags of endogenous variables as instruments. Under the assumption that past values of endogenous variable are uncorrelated with shocks Xit, any lag of an untransformed endogenous variable longer than 2 can be used as an instrument1. The same estimation technique is suitable for models with lagged dependent variables, with lags higher than the longest lag in the equation serving as instruments. Some refinements of the basic IV techniques are available for discrete endogenous explanatory variables. Efficiency is increased when fitted values from a first-stage logit or probit model for the discrete variable (with exogenous explanatory variables) are used as instruments for the second-stage model 2.6. Using the same explanatory variables directly as instruments for the discrete variable in equation 2.6, however, also produces consistent estimates. Previous empirical studies of the effects of exchange rate regimes use a number of empirical techniques in estimating the effects of regimes on macroeconomic performance. GGW(2002) use IV estimation in their basic specification, treating the regime choice as exogenously determined. Other potentially endogenous variables are instrumented by their lags. In checking the robustness of this specification, the GGW estimate a probit model for the choice of exchange rate regime in the first step and use fitted values from this model in the inflation equation to correct for the potential endogeneity of the regime choice. In both specifications no attempt is made to eliminate unobservable country specific effects. The regression with country specific effects in GGW (2002), in turn, does not correct for the potential endogeneity of the regime choice and other variables. 6.2 GDP growth The impact of exchange rate regimes on growth is ambiguous. In a simple growth accounting approach the exchange rate regime may influence economic growth either through the rate of factor accumulation (investment or employment growth) or through the growth rate of total factor productivity. The link between a pegged exchange rate and investment is not clear. A pegged exchange rate reduces policy uncertainty, interest rate volatility and real exchange rate volatility, leading to higher investment. On the other hand, a pegged exchange rate can exacerbate protectionist pressures and, if foreign trade is associated with higher productivity, it will then also reduce the efficiency of the existing capital stock. In addition, pegged exchange rates may lead to misalignments of real exchange rates and prevent efficient allocation of resources across sectors. Following GGW, the impact of exchange rate regimes on output will be analyzed based on the regression of real per capita GDP growth at constant international prices on a range of explanatory variables common in the literature. The final equation used in the estimation is as follows:

24

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

'y PC

E 0  E ExrR ExrR  E inv 'InvGDP  E Open Open 

E School School  E ToT ToT  E GovGDP GovGDP  E TaxGDPTaxGDP  E yrel ( y0 / y0US )  E dpop 'Pop  E pop log( Pop)  H (2.7) We include changes in the ratio of investment to GDP ('InvGDP) and the average number of years of schooling of the population (School). We also include a gap between the country’s per capita GDP to that of the US (y0 / y0US), to capture the relative convergence hypothesis. In addition, we include openness to trade, the tax to GDP (TaxGDP) and government balance to GDP ratios, the terms of trade, population size (Pop), and population growth ('Pop). We also include regime dummies, time dummies and income level dummies. 7. RESULTS 7.1 Inflation performance Table 2.4 reports results from the basic specification of the inflation equation described in the previous section for the ‘official’ classification in the full sample of countries, and in the sample excluding the ‘freely falling’ cases and the residual category. The full sample results confirm those obtained in previous studies. The pegged exchange rate regimes, even broadly defined, tend to have lower inflation compared to free floats. Since this result is conditional on money growth, pegging leads to a lower velocity, which may be the effect of higher credibility and lower inflationary expectations. These results are challenged when the ‘freely falling’ and the residual categories are excluded from the sample. Then the level of inflation in broadly defined pegged regimes is not significantly lower than in the floating regimes. Hard pegging, however, is still associated with lower inflation. The intermediate regimes, which perform similarly to floating regimes in the full sample results, have a significantly worse inflation record in the limited sample. Exclusion of outliers is therefore important for correct inference, especially if the focus is on low- and medium-inflation countries, as in this study. Other explanatory variables in the regression have the correct signs and most of them are statistically significant. In the same specification with the ‘natural’ classification (Table 2.5), the regime dummies are not significantly different of zero, with the exception of the ‘freely falling’ category. The dummies for aggregate regimes continue to be insignificant after this category, and the unclassified regimes, are excluded. In the results with the ‘fine’ regime classification dummies, countries with no separate legal tender have lower inflation, while some of the intermediate regimes perform worse in terms of inflation than floating regimes. The results from the two classifications are consistent with each other, although the results obtained with the RR classification are less precise than in the case of the GGW classification.

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

25

The results with money growth excluded from the set of explanatory variables (Tables 2.6 and 2.7) lead to similar conclusions. The hard pegs are associated with lower inflation than the floating regimes, with intermediate regimes performing worse. The coefficients of the regime dummies are now higher in absolute terms than in the regression with money growth, since in this case the dummies reflect not only credibility effects, but also a direct effect of lower money creation on inflation. In order to deal with the endogeneity problem, the revised, dynamic specification discussed above is estimated. Since there are not enough regime changes in the sample to estimate the models using the ‘fine’ classification, only estimates using broadly defined pegs are presented in Table 2.8. The results suggest that introduction of a de jure pegged regime is associated with inflation reduction, while exiting from the peg increases inflation. The dummy associated with a de jure pegged regime in the broad classification was not statistically significant in the base estimation in Table 2.4. This points to the importance of accounting for the endogeneity of regime choice before drawing any conclusions regarding the impact of the exchange rate regime on economic variables. The signs of coefficients of other determinants of inflation remain correct, and the variables are mostly significant. It is important to reiterate that, while these results control for the potential correlation between individual effects and the choice of regime, they do not eliminate the effect of other shocks on the joint behavior of inflation and regime choice. Exiting from the peg may be associated with a crisis related to an exogenous shock, and in this case the relationship estimated between the exchange rate regime and inflation will be spurious. 7.2 Growth performance Tables 2.9 and 2.10 report the results of the estimation of equation 2.7. With the exception of average years of schooling and terms of trade developments, all variables are statistically significant and have the expected signs. A higher investment to GDP ratio is associated with faster per capita GDP growth. This is also the case with greater openness to trade. Countries that run a budget surplus tend to grow faster, as do the countries with larger populations. Our results also support the conditional convergence hypothesis, as a higher initial income level is associated with lower GDP growth. A higher share of taxes in GDP results in slower GDP growth. Our results indicate that there are no statistically significant differences in GDP performance between pegged and intermediate exchange rate regimes. Disagregation into six categories of de jure exchange rate regimes still shows no differences in GDP growth across groups. When the finer de jure classification is employed, the only statistically significant dummy applies to countries with unclassified rule based intervention, where GDP performance is slightly worse than in case of pure floating regimes. All explanatory variables are robust to the introduction of a different set of regime dummies. Our results differ from those obtained by GGW (2002) in a similar specification, who found that GDP growth is faster under intermediate regimes by 0.7 percentage

26

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

points. The impact of pegged regimes, although positive, is not statistically significant, which is in agreement with our estimates. When the broad de jure classification is employed, GGW find that countries with hard pegs and traditional single currency pegs do not experience slower GDP growth, but GDP growth is faster under various intermediate regimes by about 1 percentage point per year as compared to free floating. Estimation of a similar specification with exchange rate regimes identified by the ‘natural’ classification indicates that there are no statistically significant differences in GDP growth under different exchange rate arrangements. All the remaining explanatory variables are robust to the introduction of a different set of regime dummies. Overall, our results indicate that controlling for other determinants, there are no significant differences in the growth performances of countries under various exchange rate arrangements. This is a similar result to that obtained by GGW, who conclude that differences in growth performance are rather small. 7.3 Output volatility The theoretical literature suggests that output volatility tends to be higher under pegged regimes if the economy is subject to real shocks and if nominal rigidities are present in the economy. When studying the conditional effect of the exchange rate regime on output volatility, we follow GGW by employing volatility of terms of trade and of the investment ratio in a specification similar to equation 2.7. Additional explanatory variables are the same as in the growth regressions. Results presented in Tables 2.11 and 2.12 suggest that both terms of trade variability and investment to GDP variability contribute to the volatility of output. The results also indicate that pegged regimes are associated with greater output volatility of about 1.5 percentage points per year. The volatility of output decreases with average income, as high-income countries experience the lowest output volatility. The ‘fine’ de jure classification indicates that the highest output volatility is recorded in countries with basket pegs and hard pegs, exceeding output volatility under floating regimes by 2.4 and 1.5 percentage points respectively. Application of the ‘fine’ classification shows that countries with secret basket pegs, currency boards, managed floats with heavy intervention, dollarization, crawling pegs, published basket pegs and monetary union membership are also associated with greater volatility of output than the floating exchange rate regime. In the ‘natural’ classification the signs are reversed. Pegged regimes (especially de facto pegs and de facto crawling bands narrower than or equal to +/-2%) are associated with significantly lower output volatility. This is a similar result to that obtained by GGW in the results based on their ‘consensus’ classification (dropping cases where the actual exchange rate behavior is markedly different from the official classification). The authors suggest that this may be due to the circumstances of particular countries in the sample, and not a reflection of the impact of the exchange rate regime on output volatility. The ‘consensus’ classification includes under the float category all countries that experience significant exchange rate movements,

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

27

among which some countries experiencing severe economic turmoil, while it excludes pegs with similarly pronounced movements in economic activity. The ‘natural’ classification of RR is likely to classify countries experiencing large shocks as floaters, and countries not subject to shocks as pegged regimes. The resulting endogeneity of the regime classification leads to potentially spurious results. First differencing – which may be successful in dealing with endogeneity when the source of the problem lies in the correlation between the regime choice and the individual effects – is likely to exacerbate the problem in the case of output volatility, since the endogeneity is more likely to stem from the correlation between regime choice and other shocks. Overall, our results are similar to those reported by GGW, although the magnitudes of the estimated impact of exchange rate regime on output volatility are significantly higher. GGW find that pegged exchange rates are associated with greater output volatility of about one-third to one-half percentage points, while we find that hard pegs have volatility that is 1.5 percentage points higher and ‘basket pegs’ volatility that is 2.4 percentage points higher than free floaters. 8. CONCLUSIONS Inflation and GDP growth performances differ significantly across countries with various exchange rate arrangements. The econometric results point to a significant role for hard pegs in improving inflation performance, and a detrimental effect of intermediate regimes on inflation. In modeling the relationship between the exchange rate choice and inflation, it is critically important to eliminate highinflation countries from the sample. The presence of outliers produces spurious estimates, pointing to a significant impact of any form of pegging on inflation performance. The choice of exchange rate regime classification seems to be less critical for the results, although using the ‘natural’ classification based on actual exchange rate behavior leads to less precise estimates. In general, it is difficult to establish a casual link between regime choice and inflation performance due to a potential endogeneity problem inherent in the regression, which is likely to be more pronounced when the ‘natural’ classification is used. This is confirmed by our estimates correcting for endogeneity of regime choice, which show that only de jure pegged regimes are associated with lower inflation. Similarly to other studies, we did not find any significant relationship between exchange rate regimes and per capita growth. The choice of exchange rate regimes is, however, significantly correlated with output volatility, leading to a potential trade-off between inflation and output performance. Flexible exchange rate arrangements are, on average, able to insulate economies from certain shocks. This insulating property comes, however, at the cost of lower credibility and therefore higher inflation. This result is not robust to the regime classification: the pegged regimes defined by the ‘natural’ classification are associated with lower output volatility. The endogeneity of the ‘natural’ classification is the most likely factor leading to this result.

28

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

NOTES 1

First lags are not valid instruments since they are correlated with first differences of shocks Xit.

REFERENCES Caballero, R.J. & Krishnamnurthy A. (2002). Excessive dollar debt: financial development and underinsurance. MIT Department of Economics Working Paper, 02-15, March 1, http://ssrn.com/abstract=306360 Calvo, G. & Reinhart C. M. (2000). Fear of floating. NBER Working Paper 7993. National Bureau of Economic Research, Cambridge MA, November. Canzoneri, M.B, Cumby R.E. & Diba, B.T. (1998). Fiscal discipline and exchange rate regimes. CEPR Discussion Paper, 1899, London. Cukierman, A. (1992). Central bank strategy, credibility and independence: Theory and Evidence. MIT Press. Fischer, S. (1993). Role of macroeconomic factors in growth. Journal of Monetary Economics, 32, 485512. Fleming, M. (1962). Domestic financial policies under fixed and floating exchange rates. IMF Staff Papers, 9, 369-380. Ghosh, A., Gulde A.M. & H. Wolfe (2002). Exchange rate regimes: choices and consequences. MIT Press. Hausmann, R., Panizza U. & Stein E. (2001). Why do countries float the way they float? Journal of Development Economics, 66 (2), 387-414. Jeanne, O. (2002). Why do emerging economies borrow in foreign currency? Presented at the conference ‘Currency and Maturity Matchmaking: Redeeming Debt from Original Sin’, IADB, Washington DC, November. Mundell, R.A. (1961). A theory of optimum currency areas. American Economic Review, 51 (3), 657-665. Mundell, R.A. (1963). Capital mobility and stabilization policy under fixed and flexible exchange rates. The Canadian Journal of Economics and Political Science, XXIX (4), 475-485, November. Romer, D. (1993). Openness and inflation: theory and evidence. Quarterly Journal of Economics, CVIII, 869-903. Reinhart, C.M. & Rogoff K.S. (2002). The modern history of exchange rate arrangements: a reinterpretation. NBER Working Paper, 8963, National Bureau of Economic Research, Cambridge MA, June. Tornell, A. & Velasco A. (2000). Fixed versus flexible exchange rates: which provides more fiscal discipline. Journal of Monetary Economic, 45, 399-436.

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

29

APPENDIX 2.1. TABLES

Table 2.1. Average inflation rates in countries grouped according to the RR classification RR classification

Average Variance

No separate legal tender Pre announced peg or currency board arrangement Pre announced horizontal band that is narrower than or equal to +/-2% De facto peg Pre announced crawling peg Pre announced crawling band that is narrower than or equal to +/-2% De factor crawling peg De facto crawling band that is narrower than or equal to +/-2% Pre announced crawling band that is wider than or equal to +/-2% De facto crawling band that is narrower than or equal to +/-5% Moving band that is narrower than or equal to +/-2% Managed floating Freely floating Freely falling Dual market in which parallel market data is missing. TOTAL

4.9 9.7

97.0 6651.6

No. of observations 130 1608

10.0

85.8

6

5.3 55.5

34.1 1341.3

361 16

19.7

594.2

24

8.7

71.7

235

8.4

47.7

759

13.2

56.7

10

11.7

133.5

462

6.5

42.9

48

16.9 8.4 302.5

3391.3 428 100.4 153 1193477.3 389

8.9

138.7

Source: Reinhart and Rogoff (2002); authors’ own estimations.

96 4725

30

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

Table 2.2. Average annual GDP growth in countries grouped according to the RR classification RR classification

Average Variance

No separate legal tender Pre announced peg or currency board arrangement Pre announced horizontal band that is narrower than or equal to +/-2% De facto peg Pre announced crawling peg Pre announced crawling band that is narrower than or equal to +/-2% De factor crawling peg De facto crawling band that is narrower than or equal to +/-2% Pre announced crawling band that is wider than or equal to +/-2% De facto crawling band that is narrower than or equal to +/-5% Moving band that is narrower than or equal to +/-2% Managed floating Freely floating Freely falling Dual market in which parallel market data is missing. TOTAL

4.9

44.8

No. of observations 122

5.1

144.8

1088

-0.6

1.3

2

3.9 3.4

12.9 20.5

283 14

4.1

10.2

24

3.7

45.3

192

4.6

18.9

593

4.5

21.9

5

5.1

20.5

286

5.4

24.7

48

3.4 3.0 0.9

31.5 10.5 41.3

341 120 340

7.0

300.8

49

Source: Reinhart and Rogoff (2002); authors’ own estimations.

3507

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

Table 2.3. Inflation and growth performance according to the ‘coarse’ RR classification Regime Pegged Intermediate Floating Freely floating Freely falling Dual market in which parallel market data is missing. Source: Reinhart and Rogoff (2002); authors’ own estimations.

Inflation 8.6 9.5 13.8 8.4 302.5 8.9

GDP growth 4.8 4.4 4.3 3.0 0.9 7.0

31

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

32

Table 2.4. Level of inflation: ‘official’ (GGW) classification with various samples

Constant 3-year SD Investment to GDP ratio Openess Avg yrs of schooling 3-year SD ToT Tax to GDP ratio Initial income/US income Population size Population growth Gov. balance Upper Income Upper Middle Income Lower Middle Income Pegged Intermediate Floating No separate legal tender Pre announced peg or currency board arrangement Pre announced horizontal band that is narrower than or equal to +/-2% De facto peg Pre announced crawling peg Pre announced crawling band that is narrower than or equal to +/-2% De factor crawling peg De facto crawling band that is narrower than or equal to +/-2% Pre announced crawling band that is wider than or equal to +/-2% De facto crawling band that is narrower than or equal to +/-5%

Coeff. 0.010 0.311 0.023 -0.002 0.098 -0.021 0.089 -0.001 -0.525 -0.251 -0.050 -0.024 -0.010 -0.008 -0.004 -0.003

(t-value) (1.140) (4.410) (8.970) (-2.480) (8.910) (-1.920) (15.500) (-1.900) (-9.300) (-16.800) (-8.840) (-5.690) (-3.130) (-1.780) (-0.987) (-0.668)

(t-value) (1.050) (4.560) (8.770) (-2.650) (8.440) (-1.930) (14.300) (-1.960) (-9.280) (-17.000) (-7.100) (-4.620) (-2.580)

-0.012 (-1.490) -0.006 (-1.090) -0.007 (-0.398) -0.012 (-2.200) 0.001 (0.039) -0.003 (-0.313) 0.005 (0.933) -0.009 (-1.850) 0.006 (0.372) 0.000 (-0.020)

Moving band that is narrower than or equal to +/-2% (i.e., allows for both appreciation and depreciation over time) Managed floating R squared No. of obs.

Coeff. 0.009 0.323 0.023 -0.002 0.094 -0.022 0.085 -0.002 -0.525 -0.255 -0.043 -0.021 -0.009

-0.009 (-1.090) -0.004 (-0.787) 0.366 1782

0.372 1782

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

33

Table 2.5. Level of inflation: the ‘natural’ (RR) classification with various samples Free Falling and Unclassified Incl. Coeff. (t-value) Coeff. (t-value) 0.003 (0.102) 0.001 (0.016) 0.472 (37.100) 0.471 (36.800) -0.687 (-8.800) -0.684 (-8.760) -0.104 (-3.070) -0.105 (-3.080) 0.083 (4.350) 0.081 (4.170) 0.046 (0.759) 0.044 (0.729) -0.007 (-1.290) -0.007 (-1.250) -0.024 (-1.750) -0.020 (-1.370) 0.013 (0.925) 0.020 (1.330) -0.011 (-0.809) -0.008 (-0.608) 0.000 (-0.011) -0.007 (-0.316) 0.004 (0.154) 0.346 (13.600)

Free Falling and Unclassified Excl. Coeff. (t-value) Coeff. (t-value) 0.016 (1.410) 0.019 (1.510) 0.504 (29.100) 0.508 (29.400) -0.253 (-7.250) -0.247 (-7.070) -0.093 (-6.230) -0.092 (-6.190) 0.029 (3.450) 0.033 (3.870) -0.031 (-1.270) -0.027 (-1.130) -0.003 (-1.270) -0.003 (-1.280) -0.031 (-5.450) -0.029 (-4.900) -0.022 (-3.480) -0.018 (-2.820) -0.024 (-4.240) -0.022 (-3.910) -0.004 (-0.437) -0.008 (-0.938) 0.004 (0.427)

-0.058 (-1.290)

-0.043 (-2.570)

0.008 (0.306)

0.001 (0.082)

-0.028 (-0.177) -0.004 (-0.167) 0.032 (0.455)

-0.001 (-0.011) -0.004 (-0.441) 0.063 (2.300)

0.044 (0.796) -0.013 (-0.494)

0.058 (2.810) -0.014 (-1.420)

-0.007 (-0.323)

-0.009 (-1.030)

-0.037 (-0.458)

-0.012 (-0.387)

-0.004 (-0.151)

0.000 (0.049)

-0.039 (-0.874) 0.016 (0.653) 0.347 (13.500)

-0.023 (-1.350) 0.011 (1.260)

Constant Broad money growth Real GDP growth ToT Growth CB Turnover Gov. balance Openness Upper Income Upper Middle Income Lower Middle Income Pegged Intermediate Floating Freely falling Dual market in which parallel market data is missing -0.009 (-0.249) No separate legal tender Pre announced peg or currency board arrangement Pre announced horizontal band that is narrower than or equal to +/-2% De facto peg Pre announced crawling peg Pre announced crawling band that is narrower than or equal to +/-2% De factor crawling peg De facto crawling band that is narrower than or equal to +/-2% Pre announced crawling band that is wider than or equal to +/-2% De facto crawling band that is narrower than or equal to +/-5% Moving band that is narrower than or equal to +/-2% (i.e., allows for both appreciation and depreciation over time) Managed floating Freely falling Dual market in which parallel market data is missing. 0.634 R squared 2347 No. of obs.

-0.008 (-0.226) 0.636 2347

0.441 1979

0.451 1979

34

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

Table 2.6. Level of inflation: the ‘official’ (GGW) classification with money growth excluded

Constant Broad money growth Real GDP gr owth ToT Growth CB Turnover Gov. balance Openness Upper Income Upper Middle Income Lower Middle Income Pegged regimes Intermediate regimes Hard pegs Single currency pegs Basket pegs Floats with rule-based intervention Floats with discretionar y intervention Dollarized Currency boar d Monetary union to outside (CFA) or inside (EMU) set of countries Single currency peg Published basket peg (SDR or non-SDR) Secret basket peg Cooperative system (EMS or predecessor) Crawling peg Target zone Unclassified rule-based intervention Managed float with heavy intervention Unclassified manage d float Other floats Float with light intervention R squar ed No. of obs.

Coeff. (t-value) 0.091 (6.890)

Coeff. (t-value) 0.091 (6.850)

Coeff. (t-value) 0.083 (6.230)

-0.147 -0.050 0.060 0.001 -0.005 -0.076 -0.016 -0.029 -0.018 0.017

-0.144 -0.047 0.062 -0.001 -0.004 -0.077 -0.012 -0.026

(-3.380) (-2.620) (6.010) (-0.028) (-1.400) (-10.600) (-1.560) (-3.800)

-0.151 -0.048 0.047 -0.027 -0.006 -0.070 -0.016 -0.031

-0.046 -0.022 -0.010 0.018 0.015

(-4.120) (-2.570) (-1.320) (2.240) (1.870)

(-3.450) (-2.740) (5.900) (0.042) (-1.830) (-11.000) (-2.180) (-4.140) (-2.620) (2.490)

(-3.620) (-2.710) (4.640) (-0.899) (-1.970) (-9.730) (-2.080) (-4.560)

-0.089 (-4.310) -0.021 (-1.740) 0.018 -0.014 0.008 -0.018 -0.012 0.098 0.093 0.081 0.102 0.013 0.009 0.024 0.165 1979

0.171 1979

(0.303) (-1.560) (0.922) (-1.950) (-1.240) (5.720) (4.380) (3.620) (5.460) (1.420) (0.609) (1.330)

0.212 1979

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

35

Table 2.7. Level of inflation: the ‘natural’ (RR) classification with money growth excluded

Constant Broad money growth Real GDP gr owth ToT Growth CB Turnover Gov. balance Openness Upper Income Upper Middle Income Lower Middle Income Pegged Intermediate Floating No separate legal tender Pr e announced peg or currency board ar rangement Pr e announced horizontal band that i s narrower than or equal to + /-2% De facto peg Pr e announced crawling peg Pr e announced crawling band that is narrower than or equal to + /-2% De factor crawling peg De facto crawling band that is nar rower than or equal to + /-2% Pr e announced crawling band that is wider than or equal to +/- 2% De facto crawling band that is nar rower than or equal to + /-5% Moving band that is narrower than or equal to + /-2% (i.e., allows for both appr eciation and depreciation over time) Managed floating R squared No. of obs.

Coeff. (t-value) 0.0831 (5.440) -0.1294 -0.0419 0.0631 -0.0144 -0.0073 -0.0657 -0.0198 -0.0284 -0.0022 0.0072 0.0209

(-3.020) (-2.290) (6.120) (-0.480) (-2.530) (-9.540) (-2.610) (-4.080) (-0.206) (0.711) (1.980)

Coeff. (t-value) 0.079 (5.280) -0.127 -0.041 0.056 -0.023 -0.006 -0.059 -0.021 -0.028

(-3.030) (-2.300) (5.520) (-0.786) (-2.060) (-8.370) (-2.640) (-4.000)

-0.060 (-2.950) 0.010 (0.860) -0.031 (-0.443) -0.007 (-0.545) 0.268 (8.400) 0.081 (3.260) -0.001 (-0.126) 0.004 (0.395) 0.039 (1.070) 0.026 (2.240)

-0.040 (-2.010) 0.030 (2.700) 0.156 1979

0.204 1979

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

36

Table 2.8. First-differences of inflation: the ‘natural’ (RR) and ‘official’ (GGW) classifications Coeff. (t-value) Constant 'Broad money growth

-0.014 0.086

Coeff. (t-value)

-3.9 -0.015

Coeff. (t-value)

-4.71 -0.001

1.32

1.037

'Real GDP growth

-0.048

-0.55 -0.043

-0.48 -0.680

'ToT Growth

-0.032

-1.52 -0.021

-1.18 -0.360

'CB Turnover

0.003

0.22

0.004

0.33 -0.010

'Gov. balance

0.070

1.68

0.070

1.81

0.29

0.007

'Openness 'Pegged (RR)

0.005 -0.047

-1.39 -0.051

'Pegged (GGW) no. of observ

0.237

0.50 -0.012

1898

-1.28 -0.001

-1.49

23.1 -6.75 -0.694

-5.29

-13.4 -0.367

-8.40

-0.603

0.063

3.80

6.17

0.215

3.19

-13.7 -0.012

-8.09

-1.89 -0.162

-5.72

-1.04 -0.053

1898

Coeff. (t-value)

1898

1898

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

37

Table 2.9. GDP growth per capita: the ‘official’ (GGW) classification

Constant Investment to GDP ra tio Openess Avg yrs of schooling ToT Tax to GDP ratio Initial income/US income Population s ize Population grow th Gov. balance Upper Income Upper Middle Income Lower Middle Income Pegged regimes Intermediate reg imes Hard p egs Single curr ency pegs Basket pegs Floats with rule-based intervention Floats with discretionary intervention Dollarized Currency boar d Monetary union to outside (CFA) or inside (EMU) set of countries Single curr ency peg Published basket peg (SDR or non-SDR) Secret basket peg Cooperative system (EMS or predecessor) Crawling peg Target zone Unclassified rule-based intervention Managed float with heavy intervention Unclassified managed float Other floats Float with light intervention R squar ed No. of obs.

Coeff. 0.0911 0.0917 0.0058 0.0008 -0.0165 -0.0328 -0.0226 0.0017 -1.1789 0.1227 -0.0002 0.0049 0.0043 0.0018 0.0016

(t-value) (4.910) (4.670) (1.930) (1.110) (-4.400) (-2.720) (-3.610) (1.950) (-19.200) (7.610) (-0.027) (1.040) (1.250) (0.486) (0.474)

Coeff. 0.097 0.085 0.006 0.001 -0.017 -0.039 -0.022 0.002 -1.176 0.120 -0.001 0.004 0.003

(t-value) (5.190) (4.290) (2.140) (1.090) (-4.510) (-3.100) (-3.550) (1.940) (-19.200) (7.420) (-0.167) (0.798) (0.757)

-0.006 0.001 0.006 0.002 0.002

(-1.180) (0.294) (1.410) (0.416) (0.550)

Coeff. 0.102 0.093 0.007 0.001 -0.017 -0.045 -0.020 0.002 -1.180 0.125 -0.003 0.002 0.003

(t-value) (5.280) (4.580) (2.120) (0.818) (-4.620) (-3.460) (-3.130) (1.810) (-19.200) (7.580) (-0.410) (0.488) (0.880)

0.002 (0.149) -0.008 (-1.020) -0.007 0.000 0.008 0.003 0.007 0.007 -0.011 -0.022 -0.002 0.001 0.008 0.000 0.279 1786

0.282 1786

(-1.190) (-0.027) (1.640) (0.696) (1.280) (0.822) (-0.971) (-1.930) (-0.231) (0.178) (0.858) (0.028)

0.286 1786

38

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI Table 2.10. GDP growth per capita: the ‘natural’ (RR) classification

Constant Investment to GDP ratio Openess Avg yrs of schooling ToT Tax to GDP ra tio Initial income/US income Populat ion size Populat ion growth Gov. balance Upper Income Upper Middle Income Lower Middle Income Pegged Intermediate Floa ting No separate leg al tender Pr e announced peg or currency board ar rangement Pr e announced hor izontal band that is nar rower than or equal to + /-2% De facto peg Pr e announced cr awling peg Pr e announced cr awling band that is nar rower than or equal to + /-2% De factor crawling peg De facto crawling band that is nar rower than or equal to + /-2% Pr e announced cr awling band that is wider tha n or equal to +/- 2% De facto crawling band that is nar rower than or equal to + /-5%

Coeff. 0.090 0.092 0.006 0.001 -0.016 -0.032 -0.023 0.002 -1.178 0.123 0.000 0.005 0.004 0.001 0.002 0.001

(t-value) (4.710) (4.660) (1.900) (1.060) (-4.340) (-2.660) (-3.720) (2.020) (-19.200) (7.560) (-0.008) (1.110) (1.220) (0.288) (0.347) (0.198)

(t-value) (4.630) (4.850) (1.880) (1.030) (-4.140) (-2.970) (-3.380) (1.740) (-19.100) (7.520) (-0.341) (0.764) (0.669)

0.003 (0.320) -0.003 (-0.487) 0.022 (1.080) 0.007 (1.220) 0.005 (0.209) -0.003 (-0.227) 0.003 (0.450) 0.001 (0.255) 0.001 (0.069) 0.003 (0.495)

Moving band that is nar rower than or equal to +/- 2% (i.e., allows for both appreciation and depreciation over time) Managed floating R squared No. of obs.

Coeff. 0.095 0.097 0.006 0.001 -0.017 -0.037 -0.022 0.002 -1.179 0.122 -0.002 0.004 0.002

0.001 (0.113) -0.001 (-0.252) 0.279 1786

0.282 1786

THE EXCHANGE RATE: SHOCK GENERATOR OR ABSORBER?

39

Table 2.11. Volatility of GDP growth per capita (3-year SD): the ‘official’ (GGW) classification

Constant 3-year SD Investment to GDP ratio Openess Avg yrs of schooling 3-year SD ToT Tax to GDP ratio Initial income/US income Population size Population growth Gov. balance Upper Income Upper Middle Income Lower Middle Income Pegged regimes Intermediate regimes Hard pegs Single currency pegs Basket pegs Floats with rule-based intervention Floats with discretionary intervention Dollarized Currency board Monetary union to outside (CFA) or inside (EMU) set of countries Single currency peg Published basket peg (SDR or non-SDR) Secret basket peg Cooperative system (EMS or predecess or) Crawling peg Tar get zone Unclassified rule-based intervention Managed float with heavy intervention Unclassified managed float Other floats Float with light intervention R squar ed No. of obs.

Coeff. -0.009 0.293 0.023 -0.001 0.099 -0.026 0.091 0.000 -0.541 -0.249 -0.048 -0.024 -0.007 0.015 0.004

(t-value) (-1.150) (4.180) (8.930) (-1.430) (9.010) (-2.410) (15.700) (0.141) (-9.660) (-16.900) (-8.310) (-5.710) (-2.330) (4.630) (1.120)

Coeff. -0.005 0.324 0.023 -0.001 0.089 -0.037 0.095 0.000 -0.537 -0.254 -0.050 -0.027 -0.007

0.015 0.004 0.024 0.002 0.006

(t-value) (-0.662) (4.650) (8.920) (-1.590) (8.080) (-3.260) (16.400) (0.192) (-9.660) (-17.300) (-8.680) (-6.360) (-2.130)

Coeff. -0.006 0.341 0.019 -0.001 0.089 -0.026 0.090 0.000 -0.519 -0.266 -0.048 -0.030 -0.009

(t-value) (-0.821) (4.890) (6.970) (-1.620) (8.050) (-2.260) (15.300) (0.316) (-9.360) (-18.000) (-8.040) (-6.670) (-2.690)

(3.440) (0.891) (6.490) (0.528) (1.760) 0.020 (2.160) 0.028 (3.850) 0.011 0.005 0.016 0.034 -0.005 0.019 0.016 0.009 0.023 0.006 0.006 0.002

0.374 1782

0.387 1782

(2.080) (1.110) (3.570) (7.710) (-1.030) (2.580) (1.640) (0.900) (2.440) (1.510) (0.775) (0.245)

0.399 1782

40

MARYLA MALISZEWSKA AND WOJCIECH MALISZEWSKI

Table 2.12. Volatility of GDP growth per capita (3-year SD):the ‘natural’ (RR) classification

Constant 3-year SD Investment to GDP ratio Openess Avg yrs of schooling 3-year SD ToT Tax to GDP ra tio Initial income/US income Population size Population growth Gov. balance Upper Income Upper Middle Income Lower Middle Income Pegged Intermediate Float ing No separate legal tender Pre announced peg or currency board arrangement Pre announced horizontal band t hat is narrower than or equal to + /-2% De facto peg Pre announced crawling peg Pr e announced cr awling band that is narrower than or equal to + /-2% De factor crawling peg De facto crawling band that is nar rower than or equal to + /-2% Pr e announced crawling band that is wider than or equal to + /-2% De facto crawling band that is narrower than or equal to + /-5%

Coeff. 0.010 0.311 0.023 -0.002 0.098 -0.021 0.089 -0.001 -0.525 -0.251 -0.050 -0.024 -0.010 -0.008 -0.004 -0.003

(t-value) (1.140) (4.410) (8.970) (-2.480) (8.910) (-1.920) (15.500) (-1.900) (-9.300) (-16.800) (-8.840) (-5.690) (-3.130) (-1.780) (-0.987) (-0.668)

(t-value) (1.050) (4.560) (8.770) (-2.650) (8.440) (-1.930) (14.300) (-1.960) (-9.280) (-17.000) (-7.100) (-4.620) (-2.580)

-0.012 (-1.490) -0.006 (-1.090) -0.007 (-0.398) -0.012 (-2.200) 0.001 (0.039) -0.003 (-0.313) 0.005 (0.933) -0.009 (-1.850) 0.006 (0.372) 0.000 (-0.020)

Moving band that is nar rower than or equal to + /-2% (i.e., allows for both appr eciation and depreciation over time) Managed floating R squared No. of obs.

Coeff. 0.009 0.323 0.023 -0.002 0.094 -0.022 0.085 -0.002 -0.525 -0.255 -0.043 -0.021 -0.009

-0.009 (-1.090) -0.004 (-0.787) 0.366 1782

0.372 1782

MONIKA BLASZKIEWICZ-SCHWARTZMAN AND PRZEMYSLAW WOZNIAK

CHAPTER 3

DO THE NEW MEMBER STATES FIT THE OPTIMUMCURRENCY-AREA CRITERIA?

1. INTRODUCTION The question of whether particular countries fit the Optimum-Currency-Area (OCA) criteria has been the subject of a growing body of empirical literature. Recently, the issue was taken up in the context of the new EU member states (NMS), which are expected to join the Economic and Monetary Union (EMU). The OCA theory, with its many empirical operationalizations, offers a valuable and concise tool to assess the costs and benefits of this important decision. This chapter investigates the OCA-implied criteria using two different empirical approaches. First, we examine a number of conventional OCA criteria, such as the openness to trade with the EU-15, as well as business cycle co-movements between the Eurozone and the NMS. This is done using various indicators and several sample periods at both annual and quarterly frequencies to check the sensitivity of results and ensure the robustness of conclusions. The second approach involves the examination of nominal and real exchange rate (RER) volatility and pursues another important theme in empirical OCA literature. The central part of the variance approach concentrates on the investigation of unexpected variances of RER between respective NMS and the EMU members, which are treated as a group, as well as investigating the RER volatility of selected EMU members before and after 1999. 2. LITERATURE REVIEW The origins of modern OCA theory can be traced to contributions in seminal papers by Mundell (1961) and McKinnon (1963). In essence, the original concept of the OCA is based on weighing the costs and benefits of giving up exchange rate flexibility, which is understood to be an instrument for dealing with balance of payments (BoP) shocks. If, for example, demand for exports from a particular country falls, a real depreciation might be necessary to maintain the BoP equilibrium and full employment. With a fixed exchange rate, real depreciation has to be

41 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 41-61. © 2006 Springer. Printed in the Netherlands.

42

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK

effected by a reduction in money wages, which takes time and causes unemployment. Thus, it is argued that when dealing with shocks to the BoP, exchange rate depreciation and appreciation respectively can reduce the impact on unemployment and inflation, especially in the world of sticky nominal wages. Giving up this important stabilization instrument would only be justified in a homogenous environment with high factor mobility, where shocks are symmetrical and well correlated. In such an environment, the benefits of exchange rate flexibility would no longer be needed, while the benefits of monetary union (MU) could be fully taken advantage of. The original OCA theory relies heavily on the assumption of stationary expectations of the price level, interest rate and even exchange rate1 (McKinnon, 2000). Common shocks are thus to be used as a criterion for determining the size of currency areas. Several years later Mundell (1973) modified his views on OCA, dropping the assumption of stationary expectations and instead focusing on exchange rate uncertainty. While the earlier paper held that asymmetric shocks disqualify regions from being a single currency area, the later paper focused on showing how having one currency could help reduce the effects of such shocks by portfolio diversification and ‘economizing’ on foreign reserves. This modification is discussed further in a series of papers by McKinnon (recently in McKinnon, 2000), who explicitly criticizes one of the basic assumptions of Mundell’s early work, i.e. the postwar Keynesian belief that national monetary and fiscal policies could successfully fine tune aggregate demand in response to shocks. Giving up this instrument is the central argument against adopting a single currency in the original OCA theory. McKinnon (2000) observes that this opinion was shared even by monetarists such as Milton Friedman, who were fond of Mundell’s earlier case for an independent monetary policy, albeit for a somewhat different reason. They thought that the autonomy of the monetary policy could be the best safeguard of domestic price levels and that a floating exchange rate would naturally reflect the stance of domestic monetary policy2. However, the great volatility of exchange rates in the 1970s and the series of currency crises in the 1990s obviously shed new light on the assumption of benefits of floating exchange rates, thus making the original OCA theory somewhat less convincing. Since the seminal contributions of Mundell and McKinnon in the 1960s, the body of OCA literature has expanded very quickly, most recently due to the interest triggered by the creation of the EMU and its expected enlargement. The vast majority of papers attempt to verify empirically the advisability of forming new or expanding existing single currency areas. In spite of fundamental modifications to the theory by the originator himself (Mundell, 1973) and recent investigations by Ching & Devereux (2000; 2003), almost all of the empirical research is based on Mundell’s early findings (1961). Specifically, the assumption is commonly made that the asymmetry of shocks to which the country (region) is subjected is the main argument against joining MU. Thus, the smaller the asymmetry, the more appropriate the country is considered as a candidate for union. The issue of labor mobility and fiscal transfers as OCA criteria is taken up much less frequently. Frankel & Rose (1998) argue that deciding on the appropriateness of countries’

OCA CRITERIA FOR NEW MEMBER STATES

43

accession to MU based on historical data correlations constitutes a classic failure to take account of the ‘Lucas Critique’. Nevertheless, they do not reject the standard OCA criteria as irrelevant, but rather, argue that ‘a country is more likely to satisfy […] (them) ex post than ex ante’.

3. STANDARD STATIC OCA INDICATORS In an attempt to answer the question of whether the NMS can be regarded as potentially well fitting members of the Eurozone, we selected a number of the conventional criteria offered by the OCA literature. In this section we examine standard indicators often mentioned and empirically verified in the OCA literature (e.g. Fidrmuc & Schardax, 2000; Bratkowski & Rostowski, 2001; Gros & Hobza, 2003). Table 3.1. Trade with EU-15 as % of GDP, 1999-2003

Country

1999 2000 2001 2002 Imports Bulgaria 20.6 22.7 26.3 25.5 Romania 18.0 19.9 22.2 22.8 Cyprus 20.7 22.5 21.7 21.2 Czech Rep. 32.8 38.7 39.4 35.2 Estonia 43.1 51.6 43.4 42.7 Hungary 37.6 40.4 37.6 30.2 Latvia 22.3 21.6 22.4 23.3 Lithuania 21.1 20.8 23.1 24.4 Poland 19.2 18.0 16.6 17.8 Slovenia 34.6 36.2 35.2 33.8 Slovakia 29.0 30.8 35.2 34.6 EU-15 16.7 18.4 18.2 17.6 Luxembourg 47.3 54.6 56.2 58.3 Denmark 2.6 2.8 2.8 2.9 Germany 14.7 16.6 16.8 16.0 Greece 17.3 18.7 13.5 12.7 Spain 15.7 18.7 18.1 17.2 France 15.0 17.3 16.7 15.5 Ireland 30.8 33.9 32.7 29.0 Italy 12.0 13.2 13.0 12.6 Netherlands 29.5 31.2 29.1 28.6 Austria 27.3 29.6 30.8 29.9 Portugal 27.4 28.6 27.6 26.3 Finland 17.4 19.2 18.6 18.1 Sweden 2.2 2.5 2.2 2.1 UK 18.4 20.0 19.4 19.2

Trade with EU-15 as % of GDP 2003 1999 2000 2001 2002 2003 1999 Exports 26.8 16.1 19.6 20.6 20.3 21.2 36.7 24.3 15.6 17.9 19.2 20.3 20.9 33.6 19.5 4.3 3.9 4.1 4.2 4.1 25.0 35.6 33.0 38.7 40.3 37.8 39.8 65.8 41.6 33.3 47.1 41.2 36.0 37.0 76.4 31.6 39.7 45.5 43.7 36.8 37.7 77.3 24.2 14.9 15.6 14.9 15.0 16.2 37.2 24.0 14.1 16.0 18.1 18.8 16.6 35.2 19.8 12.5 13.3 13.4 14.7 17.6 31.7 34.0 28.1 29.4 29.5 28.0 27.2 62.8 34.6 30.1 34.6 36.2 36.3 39.6 59.1 17.4 17.8 19.5 19.4 18.9 18.5 34.5 56.4 56.0 63.0 64.9 65.8 64.3 103.3 2.7 2.8 3.1 3.0 3.1 3.0 5.4 16.4 16.7 18.9 19.4 19.4 19.9 31.5 14.0 5.6 5.7 4.4 4.1 4.2 22.9 17.0 12.7 14.9 14.8 14.2 13.8 28.4 15.3 14.6 16.1 15.6 14.9 14.4 29.6 22.0 50.2 52.5 51.4 48.1 37.9 81.0 12.0 12.5 13.5 13.3 12.6 11.8 24.6 27.9 44.8 50.9 48.8 46.4 45.8 74.3 31.4 23.4 25.7 27.0 27.7 28.0 50.7 24.0 17.9 18.6 18.2 17.8 17.2 45.3 18.0 21.5 24.1 21.5 20.8 19.8 38.9 2.2 2.4 2.6 2.2 2.1 2.2 4.6 17.5 17.1 19.2 18.3 17.3 14.4 35.6

2000 2001 2002 Imports + Exports 42.4 46.9 45.8 37.8 41.4 43.1 26.4 25.8 25.4 77.4 79.7 73.0 98.8 84.6 78.7 86.0 81.3 67.1 37.3 37.3 38.3 36.8 41.2 43.2 31.3 30.1 32.5 65.6 64.6 61.8 65.4 71.4 70.8 38.0 37.6 36.5 117.6 121.2 124.1 5.8 5.8 6.0 35.5 36.1 35.4 24.4 17.9 16.7 33.6 32.8 31.4 33.4 32.3 30.4 86.5 84.1 77.1 26.7 26.4 25.3 82.1 77.9 75.1 55.3 57.8 57.6 47.2 45.7 44.1 43.3 40.2 38.9 5.0 4.4 4.3 39.2 37.7 36.5

2003 48.0 45.2 23.5 75.4 78.6 69.3 40.3 40.6 37.4 61.3 74.2 35.9 120.7 5.7 36.3 18.2 30.9 29.8 59.9 23.8 73.7 59.4 41.2 37.8 4.3 31.9

Source: CANSTAT – Candidate Countries Statistical Bulletin www.insse.ro/canstat_Q1/canstat.htm and AMECO database of the EUROSTAT

44

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK

3.1. Trade with the EU In this subsection we examine the proportion of EU trade in GDP as well as the total trade of NMS. Both indicators provide a straightforward measure of the extent of trade links with the EU; a criterion for the OCA analysis that can be traced back to McKinnon (1963). Countries whose foreign trade pattern exhibit high gravity towards the Euro-dominated EU should be considered appropriate candidates for joining the MU. Table 3.1 shows imports and exports from and to EU-15 as a % of GDP for nine NMS, Bulgaria and Romania, as well as 15 old member states (OMS) for the period 1999-2003. The crucial parameter, i.e. the share of exports in GDP for the most recent available full year (2003), has been marked in bold. The main message the table conveys is the large and consistently growing importance of OMS as an export market for NMS. With the exception of Cyprus, whose exports to OMS accounts for a mere 4% of GDP, all NMS export well above 15% of their GDP to OMS, ranging from 16% for Latvia to 40% for Slovakia and the Czech Republic. This is a particularly good result when compared with the current members of the Eurozone, especially Greece (4%), Italy (12%), Spain (13%) and France (14%). Table 3.2. Trade with EU-15 as % of GDP, 1992-1995

Country

1992 1993 1994 Imports EU-15 13.4 11.9 12.8 Greece 15.6 14.8 14.5 Italy 9.6 8.8 9.9 Portugal 23.9 21.0 21.9 Spain 10.5 10.0 11.7 Belgium 40.1 36.9 37.5 Denmark 15.4 14.1 14.7 Germany 12.0 9.8 10.1 France 11.1 9.7 10.5 Ireland 28.9 25.5 27.3 Netherlands 27.1 21.7 23.2 Austria 20.0 18.1 18.9 Finland 11.5 11.9 12.7 Sweden 12.3 13.4 15.1 UK 11.7 10.5 11.8

Trade with EU-15 as % of GDP 1995 1992 1993 1994 1995 1992 1993 1994 1995 Exports Imports + Exports 13.7 13.4 12.6 13.6 14.5 26.8 24.5 26.4 28.2 15.4 6.8 5.5 5.3 5.7 22.3 20.4 19.9 21.1 11.2 8.9 9.7 10.6 12.0 18.5 18.5 20.5 23.2 23.2 15.3 14.3 15.9 17.5 39.2 35.3 37.8 40.6 12.6 7.8 8.4 10.2 11.1 18.4 18.4 22.0 23.7 39.1 40.2 39.7 41.1 41.0 80.3 76.6 78.6 80.1 15.9 17.3 15.6 15.4 15.7 32.7 29.7 30.0 31.6 10.7 13.4 11.4 11.8 12.4 25.5 21.2 21.9 23.1 11.2 11.2 10.0 10.9 11.6 22.3 19.7 21.4 22.8 27.3 40.7 41.2 45.4 47.6 69.6 66.7 72.7 74.9 23.4 30.9 28.3 29.5 30.5 58.0 49.9 52.7 53.9 20.2 15.9 14.2 14.6 15.6 35.8 32.3 33.5 35.7 7.9 14.4 15.6 16.8 10.5 25.9 27.5 29.6 18.5 17.2 13.6 14.8 15.9 18.3 25.9 28.2 31.0 35.4 12.7 10.6 9.6 11.0 12.2 22.4 20.1 22.8 24.9

Source: own calculations based on AMECO database.

Table 3.2 presents the analogous indicators for OMS for the early 1990s, when they began preparations to adopt the Euro. The share of exports to OMS as a % of GDP during 1991-1995 for Greece, Spain and Italy is much lower than the respective values for 1999-2002 for Latvia, which is the candidate with the lowest value of the indicator (except Cyprus). Portugal, the most EU oriented of all Club Med countries in terms of exports, was characterized by values in the range of 1617% during 1994-1995; very close to current values for Lithuania and Poland, but

OCA CRITERIA FOR NEW MEMBER STATES

45

less than half those for the NMS leaders: the Czech Republic, Slovakia, Hungary and Estonia. Another indicator, trade with OMS as a % of total trade, which is presented in Table 3.3, measures the extent to which countries are exposed to the effects of volatility of the common currency, i.e. it reflects vulnerability to shocks from third countries (see Bratkowski & Rostowski, 2001). For most NMS, exposure to extraEU trade is in the range of that of the old EU countries. The share of exports going to OMS ranged from 56 to 74% of total exports in 2003 (with the exception of Lithuania’s 42%) which is broadly in line with the OMS average of 67%. Thus, one can conclude that the exposure of NMS (as well as Bulgaria and Romania) to Euro volatility is not particularly different from that of OMS. Table 3.3 Trade with OMS as % of total trade, 1999-2002

Country

1999 2000 2001 2002 Imports Bulgaria 48.4 44.0 49.3 50.2 Romania 60.7 56.6 57.3 58.4 Cyprus 52.6 51.6 50.8 53.0 Czech Rep. 64.2 62.1 61.8 60.0 Estonia 65.3 62.6 56.5 57.8 Hungary 64.4 58.5 57.8 56.3 Latvia 54.5 52.5 52.6 52.9 Lithuania 46.5 43.3 44.0 44.5 Poland 65.0 61.2 61.4 61.7 Slovenia 68.9 67.8 67.6 68.0 Slovakia 51.7 48.9 49.7 50.3 EU-15 65.4 62.6 63.5 64.7 Luxembourg 72.1 70.4 71.6 72.6 Denmark 73.4 72.0 72.5 74.6 Germany 65.5 62.7 64.0 65.1 Greece 68.2 62.8 55.8 54.2 Spain 69.8 67.5 68.4 68.8 France 68.4 67.0 67.2 68.2 Ireland 62.8 63.3 66.8 67.0 Italy 64.5 59.7 60.3 61.0 Netherlands 57.1 53.1 53.6 55.0 Austria 81.7 79.5 79.8 79.9 Portugal 78.8 76.3 76.6 79.7 Finland 69.2 66.9 69.3 69.9 Sweden 71.1 68.2 69.8 70.9 UK 54.6 51.0 51.9 54.7

Trade with OMS as % of total trade 2003 1999 2000 2001 2002 2003 1999 2000 2001 2002 2003 Exports Imports + Exports 49.5 52.1 51.3 54.7 55.6 56.6 50.0 47.1 51.6 52.5 52.4 57.7 65.5 63.8 67.8 67.1 67.7 62.8 64.6 64.7 64.7 64.5 55.8 40.0 36.4 38.3 50.7 56.4 49.9 59.8 61.8 62.2 61.9 59.3 69.2 68.7 68.9 68.3 69.8 66.6 48.6 48.3 52.6 55.9 53.6 72.5 76.5 69.5 67.9 68.3 68.2 65.2 65.2 64.1 64.4 55.1 76.2 75.2 74.3 75.1 73.6 70.0 68.6 62.1 62.0 59.6 51.0 62.5 64.6 61.2 60.4 61.8 57.5 66.3 65.6 65.2 63.8 44.5 50.1 47.9 47.8 48.3 42.1 47.9 57.0 55.7 55.6 54.9 61.1 70.5 70.0 69.2 68.7 68.8 67.0 45.2 45.6 46.1 43.5 67.3 66.1 63.9 62.2 59.3 58.4 67.6 66.0 65.1 63.7 63.0 51.4 59.4 59.1 59.9 60.5 59.5 55.3 53.8 54.4 55.1 55.4 65.0 68.0 66.8 66.5 66.5 67.0 66.7 64.7 65.0 65.6 66.0 73.4 78.5 76.6 77.7 75.2 77.1 75.4 73.6 74.8 73.9 75.3 73.3 70.7 70.4 69.3 69.5 70.1 72.0 71.2 70.8 71.9 71.6 65.6 65.0 64.1 63.0 62.7 64.0 65.2 63.5 63.5 63.8 64.7 55.5 60.8 54.9 50.6 52.5 55.0 66.2 60.8 54.4 53.8 55.4 69.0 73.5 72.8 74.1 74.5 74.8 71.4 69.8 70.9 71.3 71.5 69.2 65.0 64.5 63.8 64.7 65.7 66.7 65.8 65.5 66.4 67.4 62.5 67.2 64.6 64.1 65.9 62.3 65.5 64.1 65.1 66.3 62.4 60.5 62.9 60.3 59.5 59.2 59.6 63.7 60.0 59.9 60.1 60.0 54.4 81.8 81.1 81.2 80.1 79.9 69.8 67.6 68.1 68.2 67.9 80.7 75.3 73.7 73.6 73.5 73.6 78.6 76.7 76.8 76.7 77.2 78.4 84.2 81.5 81.4 81.6 80.6 80.8 78.3 78.5 80.5 79.3 68.3 65.2 62.9 60.4 61.0 60.1 66.9 64.6 64.2 64.8 63.7 71.7 62.4 60.1 58.7 58.3 58.4 66.3 63.8 63.8 64.0 64.4 55.5 60.6 59.2 59.6 61.1 58.8 57.3 54.7 55.4 57.6 56.9

Source: CANSTAT – Candidate Countries Statistical Bulletin www.insse.ro/canstat_Q1/canstat.htm and EUROSTAT

The overall impression is that the NMS, as well as Bulgaria and Romania have successfully re-oriented their economies towards EU-15 markets. On average, their exports to the EU-15 account for a much higher share of GDP than was the case

46

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK

with the Club Med countries in the first half of the 1990s when they were preparing to join the Eurozone. Moreover, during recent years this indicator for all NMS-8 has been higher than for many EMU members. Therefore, as far as the importance of trade with the EU-15 is concerned, NMS score very well and – according to this standard OCA criterion - are expected to benefit greatly from joining the EMU. 3.2. Correlation of business cycles Correlation of real growth rates, growth rates of industrial output and changes in unemployment are used in the OCA literature to measure the co-movements of business cycles. Close correlations are interpreted as an indicator of the symmetry of shocks, and hence as a sign favoring entrance into MU. However, in the case of transition economies they must be used with caution because of the problems of interpreting output and employment movements as business cycles. The shock related to the transition from a planned to a market economy started for most of the investigated countries in the early 1990s, but when it ceased to dominate macroeconomic developments is the subject of debate. One can be sure that the data from the first half of the 1990s carry very little information about the cycle (GDP, industrial production, unemployment rate). On the one hand, they contain a great deal of noise related to structural changes in these economies and frequent changes in the law, as well as purely methodological modifications to definitions and methods of calculation of economic time series. On the other hand, the more one moves towards the end of the 1990s, the more limited the effect of the transition shock becomes and the data series contain more credible and trustworthy information. Hence in this analysis we work with the most recent data possible and check the robustness of results by examining alternative time periods. 3.2.1. Correlation at annual frequency. Table 3.4 shows the coefficients of correlation between real growth rates and annual changes in unemployment rates in the Eurozone and respective countries during the periods 1996-2004 and 1991-1995 (for the Club Med countries). For the period 1996-2004, two sub-periods were checked in order to examine the results’ sensitivity and to verify whether the business cycle correlation is tighter in the more recent period. In contrast to the case of export-based coefficients, indicators of business cycle correlation seem to be smaller for NMS than for OMS. While for the OMS (except Greece) the coefficients for real growth rates are all well above 0.53, respective figures for NMS are extremely dispersed and often take on negative values. Real growth rates in Estonia, Slovakia, Latvia and Lithuania exhibit a consistently negative correlation with Eurozone growth rates. A high positive correlation was only detected for Hungary and Slovenia, while a moderate correlation was found in Poland. For these countries, it is also very visible that the correlation gets stronger in the second, more recent period. This indicates that a process of convergence is occurring and that the paths of GDP growth in these countries are moving closer and closer to those of the EU-15. Comparison with the coefficients calculated for the

OCA CRITERIA FOR NEW MEMBER STATES

47

Club Med countries during 1991-1995 (all of which are very high) suggests that NMS are much more diverse and still subject to many idiosyncratic shocks. Much in line with previous findings, the high correlation of unemployment changes found for the Eurozone members (with the exception of Greece) is coupled with very strange patterns of correlation for the NMS. A relatively high positive correlation is detected for Hungary and Slovenia, while for the remaining countries correlation rates are very often negative and close to –1. This is in stark contrast to the correlation of unemployment rate changes for Club Med countries during 19891994, all of which are positive and very high. Table 3.4. Correlation between annual real growth rates in the Eurozone and selected countries

Country Bulgaria Romania Turkey Czech Rep. Estonia Hungary Latvia Lithuania Poland Slovakia Slovenia Greece Italy Portugal Spain Belgium Denmark Germany France Ireland Netherlands Austria Finland Sweden UK

Correlation between annual real growth rates in the Eurozone and respective countries during: 1991-1995 1996-2004 1999-2004

0.99 0.95 0.84 0.97

0.10 -0.5 -0.13 -0.28 -0.03 0.71 -0.22 -0.53 0.33 -0.61 0.75 0.03 0.91 0.71 0.91 0.80 0.84 0.98 0.95 0.76 0.84 0.83 0.67 0.82 0.87

-0.08 -0.62 -0.13 0.21 -0.17 0.95 -0.35 -0.74 0.40 -0.75 0.86 -0.03 0.92 0.76 0.92 0.88 0.98 0.99 0.99 0.85 0.94 0.91 0.81 0.84 0.91

Correlation between annual change in unemployment rates in the Eurozone and respective countries during: 1991-1995 1996-2004 1999-2004

0.73 0.89 0.87 0.97

-0.60 -0.22 0.13 -0.45 -0.53 0.78 -0.13 -0.75 -0.59 -0.84 0.45 -0.15 0.36 0.77 0.85 0.76 0.36 0.93 0.95 0.74 0.52 0.79 0.36 0.88 0.26

-0.89 -0.25 0.30 -0.37 -0.65 0.84 -0.54 -0.87 -0.87 -0.91 0.72 -0.11 0.57 0.98 0.90 0.91 0.76 0.97 0.95 0.94 0.80 0.84 0.81 0.87 0.73

Source: Authors’ calculations based on AMECO database.

Moreover, with few exceptions, unemployment rate correlation coefficients for NMS do not coincide with those of the real growth rate. This is the result of fundamental structural changes, changing laws relating to the legal status of unemployment benefits, and the evolution of the welfare state in transition economies. Therefore, although the correlation between unemployment rate changes

48

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK

is a standard criterion used in the literature to gauge the appropriateness of a country joining MU, in the case of NMS it needs to be applied with the highest degree of caution. 3.2.2. Correlation at quarterly frequency. To detect the short-term collinearity of real sector developments, as well as to check the robustness of the results obtained for annual frequency data, the same correlation indicators were calculated using quarterly data4. In addition to real growth rates and unemployment rates changes, we analyzed annual growth rates of industrial output. To check the sensitivity of results to changing the sample period, correlations were calculated for periods 1996:Q1-2004:Q3 and 1999:Q1-2004:Q3. The resulting correlation coefficients are presented in Table 3.5. The short-term correlations confirm previous findings. There are only two countries that score very high correlations in both the annual and quarterly data: Hungary and Slovenia. Poland exhibits a moderate correlation for GDP and industrial production, but a negative one for unemployment. Coefficients for all three variables for the remaining countries are unstable and often negative. All this is in stark contrast to the correlations of the Eurozone countries, the majority of which exhibit high and positive coefficients. An exception to this rule is Greece (GDP and unemployment), as well as Portugal (industrial output). Correlations between indicators of Club Med countries and Germany5 in the early 1990s (not reported in the table) point to much closer co-movements in the real economy than those detected for the NMS (correlation coefficients usually exceed 0.5, especially when the sample period includes the period 3 – 4 years prior to the Club Med countries joining the ERM). Summing up, it should be mentioned that unlike shares of EU exports in GDP, the correlation of output and unemployment movements provide a mixed picture. When checked against annual and quarterly frequency data and for various sample periods, these correlations point to substantial diversity within the NMS group. The only countries whose correlation with the EU-15 is close to that of the OMS are Hungary, Slovenia and possibly Poland. For all other countries, the evidence for comovements is very weak and sensitive to the indicator, time period or the data frequency. This performance is in stark contrast to the current performance of the OMS and to that of the Club Med countries in the first half of the 1990s. However, as the OCA theory reviewed in the second chapter makes clear, these traditional indicators have frequently been criticized for being ‘static’ and failing to account for the endogeneity in business cycle indicators. It is argued that once a country is admitted to a common currency area, business cycle correlations rise, shocks get more symmetrical and trade with the area soars (evidence for this occurring in the NMS can be found in Fidrmuc, 2001; see also Chapter 4 of this volume). Therefore, it is not appropriate to gauge countries’ readiness for joining a MU based on the ex-ante values of those indicators.

OCA CRITERIA FOR NEW MEMBER STATES

49

Table 3.5 Correlation of business cycles at quarterly frequency

Country

Correlation between annual real growth rates in the Eurozone and respective countries during:

1996:Q12004:Q3 Bulgaria 0.26 Romania -0.62 Turkey 0.03 Czech Rep. -0.28 Estonia 0.20 Hungary 0.69 Latvia 0.06 Lithuania -0.31 Poland 0.43 Slovakia -0.34 Slovenia 0.66 Greece -0.10 Italy 0.85 Portugal 0.55 Spain 0.80 Belgium 0.80 Denmark 0.70 Germany 0.96 France 0.92 Ireland 0.73 Netherlands 0.80 Austria 0.80 Finland 0.63 Sweden 0.76 UK 0.63

1999:Q12004:Q3 0.08 -0.62 -0.02 0.23 -0.02 0.90 -0.19 -0.65 0.52 -0.58 0.77 0.03 0.87 0.64 0.86 0.85 0.76 0.98 0.95 0.71 0.92 0.86 0.67 0.80 0.75

Correlation between annual growth of industrial production in the Eurozone and respective countries during: during: 1996:Q11999:Q12004:Q3 2004:Q3 0.15 0.15 -0.41 0.14 0.27 0.23 0.00 0.00 0.35 0.35 0.88 0.88 0.22 0.02 -0.28 -0.28 0.46 0.51 0.27 0.27 0.66 0.66 0.56 0.70 0.90 0.96 0.11 -0.03 0.79 0.74 0.82 0.86 0.49 0.50 0.95 0.97 0.93 0.90 0.67 0.62 0.62 0.80 0.91 0.93 0.88 0.88 0.72 0.72 0.74 0.84

Correlation between annual change in unemployment rates in the Eurozone and respective countries during: 1996:Q12004:Q3 -0.85 -0.17 0.03 -0.34 -0.53 0.70 -0.43 -0.80 -0.57 -0.81 0.49 -0.13 0.35 0.74 0.84 0.73 0.31 0.91 0.93 0.73 0.49 0.67 0.34 0.84 0.26

1999:Q12004:Q3 -0.85 -0.22 0.03 -0.34 -0.59 0.79 -0.43 -0.80 -0.82 -0.81 0.72 -0.14 0.53 0.90 0.89 0.90 0.70 0.95 0.94 0.93 0.73 0.74 0.75 0.87 0.71

Source: Authors’ calculations based on Eurostat’s METADATA and AMECO databases.

On the other hand, even if we accept this criticism of traditional OCA criteria, it still leaves room to apply them in a comparative framework with countries that were also pre-accession to the EMU. In this analysis, this has been done for the Club Med countries during the first half of the 1990s. Even if one accepts that the computed indicators are in fact endogenous and would all improve once countries joined the Eurozone, it still makes sense to assume that the higher they are prior to the EMU membership, the higher they will be after the accession. Thus, an examination of these indicators prior to adoption of the Euro gives a good estimate of where candidate countries are now and where they could be, should they join the Eurozone.

50

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK 4. THE EXCHANGE RATE VARIABILITY APPROACH TO OCA

This chapter presents an empirical application of another OCA approach, i.e. the investigation of RER volatility to measure the extent of shock asymmetry and thus coherence with the common currency area. 4.1. Asymmetric shocks and other issues We concentrate on estimates of variation of exchange rates in the NMS. This choice of methodology is governed by the fact that exchange rate variability seems to be the most comprehensive way of assessing whether joining a MU will be more or less attractive than retaining the status quo. This is chiefly because it allows for the incorporation of other factors recognized as important for the creation of OCA (for example, factor mobility, the degree of market diversification, fiscal integration, degree of openness, etc.). Nevertheless, it also has its disadvantages. For example, the variance approach ignores the fact that exchange rate fluctuations are caused by a variety of factors, making it difficult to isolate the effects of a particular event; it does not allow for separating out changes in RER caused by nominal factors, such as financial market movements, from movements caused by real factors. As shocks to domestic money supply are short lasting and can actually be better tuned once a MU is created, this concern should not be ignored. Another point of criticism is that the methodology ignores the fact that, for example, currency boards with an anchor different from the Euro naturally exhibit higher fluctuations of the Euro exchange rate. This would be an important limitation in the context of our study since some NMS have run currency boards anchored to USD while the others to DEM. However, we argue that all NMS except Latvia already have either Euro-based currency boards or Euro-dominated reference baskets (see Egert & Kierzenkowski, 2003)6. Moreover, in order to be admitted to the EMU, Latvia, like all EMU candidates, must limit fluctuations of its national currency against the Euro (+/- 15% around the central parity), and hence we are interested in fluctuations of nominal exchange rates (NER) against the Euro and not USD or any other currency. Also, in our analysis we attempt to investigate to what extent NER plays a role in adjusting RER between regions (i.e. NMS and EMU) exposed to asymmetric shocks. These caveats notwithstanding, we think that the analysis of RER variability should give a more complex and complete insight into the issue of cycle-comovements than would any static indicator or VAR-based methods7. Instead of concentrating on the potential sources of asymmetric shocks or restricting them to a one-variable level, the variability approach involves looking at the behavior of the exchange rate, which is assumed ultimately to reflect adjustments to shocks. In this context, static methods seem rather primitive as they are bound to provide a very fragmented picture of the macroeconomic conditions underlying the decision to join a MU. Moreover, their investigation ex ante risks being charged with failing to account for the ‘Lucas Critique’.

OCA CRITERIA FOR NEW MEMBER STATES

51

4.2. Nominal Exchange Rate Variability As has already been pointed out, financial shocks can be better tackled once a MU is actually created. This is because a single currency minimizes impediments to money flows across national borders (Patterson & Amati, 1998). From this perspective, and given that the NMS face increased net capital inflows (i.e. structural funds, increased creditworthiness), giving up national currencies should reduce the distress related to NER movements. However, given market rigidities, it is often the NER that adjusts in response to required changes in the RER. A stable NER is not only a pre-requisite for minimizing costs of formulating a MU, it is also required by the Treaty. Table 3.6. NER volatility: distance from the Club Med average Club Med Average 1996-1998 1993-1995

BUL

CZE

EST

HUN

LAT

LIT

POL

ROM

SLK

SLO

-0.73 -2.44

1.04 -0.67

-0.74 -2.45

1.94 0.23

0.77 -0.94

-0.36 -2.07

3.57 1.86

1.79 0.08

0.60 -1.11

-0.71 -2.42

Source: Authors’ calculation based on IMF IFS and ECB data.

Table 3.6 sets out a measure of NER stability for the NMS in terms of the difference between the average (2001-2003/2004) NER volatility for the respective NMS country vis-à-vis the Euro and the average volatility of the Club Med countries vis-à-vis the Euro over the three years preceding the introduction of the common currency (i.e. first line in Table 3.6). NER volatility in the NMS was calculated over this three-year period (November 2001 to September 2004) and was defined (in line with Gurjarati, 2003) as the mean of the squared deviation of the first difference of the logged NER from its mean. The volatility of the Club Med countries’ exchange rates was calculated over the period 1996-1998 (January 1996 to December 1998) in the same way as the volatility for the NMS. In all cases except for Bulgaria, Estonia, Lithuania and Slovenia, the differences in volatility are positive, indicating that the NMS have more volatile NERs than the ‘periphery’ EMU members. Not surprisingly, Bulgaria, Lithuania and Estonia, countries with currency board arrangements, have more stable NERs relative to the Club Med average. The stability of the Slovenian currency is clearly evident, despite its de jure managed float regime; it exhibits negative differences in volatility8. Among the countries which still have volatile exchange rates, the country with the smallest distance from the Club Med average is Slovakia, followed by Latvia, and the Czech Republic. Poland, Hungary and Romania exhibit the most volatile NER9. However, given that the two-year window for ERM participation for past candidates covers the period from March 1996 to February 1998, it is argued that for comparative analysis the early 1990s should be used as a reference point. Assuming that the NMS will join the ERM-II as soon as possible, the period 2001-2003/2004 represents approximately the same stage in their accession process as 1993-1995 for the Club Med countries.

52

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK

The second line in Table 3.6 shows the difference between the average exchange rate volatility for the NMS (calculated as above) and the average volatility of the Club Med countries between 1993 and 1995, instead of between 1996 and 1998. The results are striking. Now, only in the case of Poland is the NER significantly more volatile than the Club Med average (i.e. having a distance greater than one). Countries like Latvia, Slovakia or the Czech Republic have even smaller volatility (i.e. a distance smaller than zero); the distances for Hungary and Romania are close to zero. In conclusion, although the NMS still have quite volatile exchange rates when compared with the Club Med average of 1996-1998, this does not appear to be the case when compared with the Club Med’s 1993-1995 average. Clearly, there are four countries which outscore and three countries which underscore the rest of the sample as well as the Club Med average (irrespective of the comparable point in time). These are, respectively, Bulgaria, Estonia, Lithuania and Slovenia; and Hungary, Poland and Romania. However, the stability of Bulgarian, Lithuanian and Estonian NERs is not surprising given their currency board arrangements. 4.3. Empirical Analysis of RER Movements We turn now to the central part of our analysis, which focuses on RER fluctuations. In order to distinguish between real and nominal shocks, we work with different frequencies (i.e. monthly and quarterly data)10. We make the crucial assumption (in line with von Hagen & Neumann, 1994) that high-frequency RER changes mostly reflect nominal shocks and low-frequency RER changes are principally due to real shocks. This distinction is the basis for evaluating the differences between asymmetric real and nominal shocks. Similarly, since the real variability is influenced by nominal variability, by working with different frequencies we have tried to tackle the problem of the inability to distinguish between the two types of shocks, which is a frequently criticized shortcoming of the variance approach to assessing costs criteria from OCA theory. 4.3.1. Methodology Building on the findings set out above, we estimated unexpected (i.e. conditional) RER variances between respective NMS and EMU members treated as a group (i.e. real exchange rates were deflated by the ratio of prices between a particular NMS and Eurozone HICP inflation)11. This approach draws on Vaubel (1976) and is similar to that of von Hagen & Neumann (1994) and Gros & Hobza (2003)12. Given that our aim is mainly to address the question of whether the NMS could benefit from adopting the Euro (based on OCA cost-benefit analysis), the choice of this reference group seems to be appropriate. To facilitate assessments of the magnitude of these RER variances (i.e. to decide when the variance should be considered large and when small) estimates of the observed RER volatility of selected EMU members are also provided13. Since looking at historical Euro RER volatility (i.e. prior to the actual creation of the union) might be considered questionable, we have allowed for various sensitivity

OCA CRITERIA FOR NEW MEMBER STATES

53

checks. For example, we also look at RER volatility between a particular member state and Germany (i.e. a ‘core’ EMU member), as well as at price differentials. 4.3.2. Sample and Data In order to distinguish between different exchange rate regimes, as well as to separate out the early stages of the transition period, we have divided our sample into three parts: the mid-transition period of 1993-1995; the late transition of 19961999; and the pre-accession/accession period of 1999-2003/2004. These subsamples more or less correspond to the steps taken by some NMS in their movement towards more flexible regimes (see for example Egert & Kierzenkowski, 2003). The fact that respective NMS represent a broad range of exchange rate regimes allows us to comment on the impact of those arrangements on RER volatility. For example, we ask if Bulgaria, Estonia, Latvia and Lithuania necessarily have less volatile real exchange rates than Poland, the Czech Republic or Romania. From the perspective of the Club Med countries, as well as France and Germany, the choice of sample period was governed by two factors. Firstly, 1993 marks the end of the EMS and therefore allows for NER fluctuation within a band of +/-15%. This ensures minimum policy coordination between countries and is important for comparative purposes. Secondly, the fact that we also look at the data on past exchange rate variations and compare those with the exchange rate volatility of NMS allows us to address the question of endogeneity. For example, we look at the period when conditions for the Club Med countries were not influenced by structural changes induced by the creation of the MU itself. Data on monthly average NER against the USD, as well as consumer price indices up to September 2004 - for all countries under consideration - comes from the IMF IFS. In order to calculate exchange rates against the Euro we used the ECB reference EUR/USD exchange rate (against the ECU up to December 1998). The Eurozone price index (HICP) is also sourced from the ECB. As this series only starts in 1995, before this date it is approximated by a producer price index for the entire region. 4.3.3. Estimation Techniques. We started our estimation of the conditional standard deviations (STD) of RER shocks by computing RER for NMS. We first took the natural logarithm of normalized price and exchange rate indices with a common base in 1995:1, and then defined log RER for the respective NMS as the NER adjusted for price changes between the country in question and the Eurozone. An increase in the RER index indicates depreciation. For the Club Med countries, as well as for France and Germany, we calculated three different representations of the RER. First, up to 1998, we used the same definition as for the NMS, but we also calculated the RER with respect to DEM ( in which case Germany was dropped from the sample)14. Since NER between the EMU countries equals 1 as of 1999, we also looked at price differentials between those countries and Eurozone inflation over the full sample (rows marked ‘Union’ in Tables 3.8 and 3.10). This provided us with a rough estimate of price convergence

54

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK

after the creation of the EMU. Then, we derived the unexpected component of RER changes (i.e. fluctuations which cannot be explained by past RER movements) for each country of interest by regressing seasonally adjusted RER changes on their own lags15.

' RER i,t = b1 + b2'RER i,t-1 + b3'RER i,t-2 +……+ b12'RER i,t-12 + ui,t

(3.1)

Residuals ui,t obtained from these regressions represent conditional RER shocks (see von Hagen & Neumann, 1994). Later, our analysis involved STD of these shocks: s=[var(ui,t )]1/2

(3.2)

In some cases, the best performing equations were equations that not only contained autoregression but also a moving average component (ARMA)16. Because the unexpected component in our autoregressive model (i.e. residuals from the estimated model) is itself a generated regressor (i.e. a deviation from the mean), we also tried to instrument the conditional STD. However, the Hausman specification error test we performed did not support this method of estimation. To obtain white-noise errors (ui,t) from our model (equation 3.2) where necessary, we used dummy variables. Necessarily, this lowered computed standard errors (and hence our measure of exchange rate variability). However, events responsible for lack of normality (e.g. financial crises, random exchange rate movements, contagion effects from other markets) are generally outliers and are unlikely to recur in the future in any systematic manner. To some extent, therefore, this also corrects for the negative bias (i.e. bias due to speculative pressures or irresponsible central bankers) in OCA suitability estimates. In order to check whether volatility changes in RER are significant (i.e. test for variance equality between sub-samples), we performed various statistical tests. Von Hagen & Neumann (1994) propose White’s tests for heteroskedasticity. In addition, we carried out an ARCH test, as we believe that financial market data often follow an ARCH process. Where this was the case, the presented standard errors are errors from the mean equation of an ARCH model17. Finally, unlike von Hagen & Neumann (1994), we did not use interactive dummies on the lag terms from the autoregressive model (equation 3.2) in order to allow for structural breaks, arguing that it is inappropriate to pool regressions for which variances are believed to be different (i.e. stability tests based on dummy variables or pooled regressions explicitly assume equal variances). Therefore, in order to obtain conditional variances, we decided to estimate separate regressions for each sub-sample.

OCA CRITERIA FOR NEW MEMBER STATES

55

4.3.4. Results Table 3.7 summarizes the estimates of conditional STD of monthly RER shocks for the NMS. Among them, there are three countries for which STD of RER shocks show a consistent and decreasing trend. These are Estonia, Romania and Slovenia. Hungary, Latvia, Lithuania and Poland managed to decrease the variance of RER shocks between the II and I sub-sample (although in Hungary the change was statistically insignificant); in the III sub-sample real exchange rates again became more volatile. In Bulgaria there is clear evidence of stabilizing policies between 1998 and 2003/2004. The same is true for the Czech Republic. As for Slovakia, there was a continuous increase in RER volatility throughout the whole estimation period. Table 3.7. Short-run (monthly data) volatility (NMS) I 93-95 BUL No*/Yes* 3.47 CZE No/Yes* 0.67 EST Yes*/Yes* 1.13 HUN Yes/No* 1.63 LAT Yes*/No* 2.40 LIT Yes*/No* 2.75 POL Yes*/No 1.85 ROM Yes*/Yes 3.69 SLK Yes/No* 0.71 SLO Yes/Yes* 0.78 Average 1.91 Average (excl. BUL and ROM) 1.49 Volatility changes

II 96-98 12.51 1.59 0.88 1.07 1.13 1.78 1.68 3.56 1.13 0.76 2.61 1.25

III 99-04 1.17 1.35 0.37 1.47 1.50 2.03 1.97 2.27 1.45 0.40 1.40 1.32

White Heteroskedasticity 93-98 96-04 0.82 0.00 0.20 0.00 0.00 0.00 0.76 0.30 0.00 0.26 0.00 0.00 0.03 0.90 0.89 0.32 0.65 0.50 0.83 0.00

ARCH 93-98 96-04 0.00 0.00 0.58 0.57 0.07 0.07 0.96 0.01 0.00 0.15 0.16 0.00 0.09 0.56 0.06 0.24 0.25 0.08 0.71 0.06

Notes: Yes-convergence, No-divergence, i.e. a decrease/increase in STD of RER between the two tested sub-samples (I-II and II-III); * - Statistically significant changes in STD of RER between the two sub-samples (based on White Heteroskedasticity and Autoregressive Conditional Heteroscedastic (ARCH) errors tests and 15% significance levels). If the null hypothesis is rejected then errors are heteroskedastic, i.e. the changes in conditional RER variances between sub-samples are statistically significant. Columns from 6 to 9 report P-values of statistical tests conducted. Source: Authors’ calculation based on IMF IFS, and ECB data

As in the case of NER volatility, the broad range of exchange rate regimes adopted by NMS seems not to matter for RER stability. The hypothesis that less flexible regimes contribute to more stable RER was not confirmed by the data (as illustrated by comparisons of Poland and Lithuania, Hungary and the Czech Republic, Poland and the Czech Republic). This fact can be interpreted to mean that NER flexibility is not necessary to accommodate RER shocks. If we compare the RER shocks in the NMS with the Club Med average in the early 1990s, as well as in the years preceding the launch of the EMU, RER volatility in the NMS is, on average, over two times higher than that in the Club Med countries in 1996-1998 and 1.2 times the variance of the Club Med in the early 1990s.

56

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK

In Table 3.8 we present the detailed results for the selected EMU members. The data presented here is, as in the case of the NMS, for estimated conditional standard errors of real exchange rates shocks. In the same table, lines marked ‘Union’ are STD of residuals from the regressions (equation 3.2) where real exchange rates were calculated on the assumption that NERs are equal to one (i.e. as the price differential between the respective EMU members and the whole EMU). Table 3.8. Short-run (monthly data) volatility (member states) Volatility changes Germany Yes* Union Yes*/No* France Yes* Union Yes*/No Italy Yes* Union Yes/No Greece No Union Yes*/ Yes Portugal Yes* Union Yes*/ No* Spain Yes* Union Yes/No* Average (Club Med) Average (Union)

I 93-95 0.69 0.23 0.52 0.17 1.98 0.43 0.44 0.52 1.21 0.58 1.20 0.18 1.21 0.43

II 96-98 0.60 0.14 0.30 0.10 0.79 0.25 0.66 0.47 0.73 0.19 0.50 0.13 0.67 0.26

III White Heteroskedasticity 99-03 93-98 96-03 0.15 0.16 MEAN EQ 0.05 MEAN EQ

0.14 0.00 0.00 0.27 0.55 0.95 0.26 MEAN EQ 0.62 0.26 0.15 0.54 0.25 0.70

0.77 0.77 0.80 0.46 0.17

ARCH 93-98 96-03 0.33 0.09 0.02 0.08 0.06 0.77 0.45 0.16 0.97 0.64 0.00 0.70 0.04 0.54 0.09 0.03 0.38 0.00

0.26

Notes: see notes to Table 3.7 Source: Authors’ calculation based on IMF IFS, OECD and ECB data

Table 3.8 shows unambiguously that all countries except for Greece intensified their efforts to lower RER volatility at the onset of the introduction of the Euro . Moreover, in all countries the reduction in volatility was found to be statistically significant. If we compare the magnitude of conditional variance of price differentials with that of RER before the Eurozone was actually launched, it is clear that the NER played a destabilizing, rather than stabilizing, role in all countries under consideration. Nevertheless, in all cases except Greece, price convergence has been less clear. Despite convergence in the years 1993-1998, we fail to report further price convergence once the EMU was formed. 5. QUARTERLY VOLATILITY CHANGES Now we turn to estimates of conditional STD of relative RER changes obtained for lower frequency (quarterly) data18. Since the real variability of exchange rates is influenced by nominal variability, by working with different frequencies we attempt to eliminate the problem of nominal variability in RER movements. This distinction also serves as a basis for evaluating the differences between asymmetric real and nominal RER shocks.

OCA CRITERIA FOR NEW MEMBER STATES

57

Table 3.9. Long-run (quarterly data normalized to monthly) volatility (NMS) I 93-95 BUL No*/Yes* 2.57 CZE No/No* 0.76 EST Yes/Yes* 1.32 HUN Yes/No 1.33 LAT Yes*/No 2.65 LIT Yes*/Yes 3.52 POL Yes/No 0.80 ROM No/Yes 5.01 SLK No/Yes 0.19 SLO No/Yes* 0.49 Average 1.87 Average (excl. BUL and ROM) 1.11 Volatility changes

II 96-98 12.51 1.16 0.84 1.04 1.40 2.33 1.23 5.04 0.83 1.16 2.75 0.78

III 99-04 0.94 1.50 0.39 1.26 1.71 1.77 2.84 2.25 1.43 0.25 1.43 0.96

White Heteroskedasticity 93-98 96-04 0.80 0.00 0.85 0.19 0.71 0.15 MEAN EQ 0.93 0.03 0.85 0.09 MEAN EQ MEAN EQ 0.89 MEAN EQ 0.04 0.51 0.31 0.52 0.02

ARCH 93-98 96-04 0.06 0.02 0.78 0.43 0.50 0.92 0.71 0.48 0.03 0.46 0.49 0.98 0.85 0.40 0.39 0.31 0.78 0.80 0.32 0.36

Notes: see notes to Table 3.7 Source: Authors’ calculation based on IMF IFS, OECD and ECB data

Table 3.10. Long-run (quarterly data normalized to monthly) volatility (member states) Volatility changes Germany Yes* Union Yes/No France Yes Union Yes*/No Italy Yes* Union Yes/Yes Greece No Union Yes/Yes Portugal Yes Union Yes/No Spain Yes* Union No/No Average (Club Med) Average (Union)

I 93-95 0.64 0.21 0.37 0.16 1.88 0.36 0.15 0.54 1.00 0.41 1.43 0.30 1.11 0.40

II 96-98 0.38 0.10 0.29 0.04 0.90 0.28 1.14 1.14 0.43 0.18 0.33 0.14 0.70 0.44

III White Heteroskedasticity 99-03 96-98 96-03 0.81 0.11 0.99 0.35 MEAN EQ

0.09 0.18

0.90

MEAN EQ

0.18 0.82 0.68 0.18 0.74

0.92 0.92

MEAN EQ

0.21 0.48 0.00 0.12 0.80

MEAN EQ

0.24

ARCH 96-98 96-03 0.12 0.70 0.88 0.70 0.10 0.71 0.05 0.86 0.54 0.74 0.12 0.54 0.61 0.49 0.69 0.10 0.37 0.58

0.17

Notes: see notes to Table 3.7 Source: Authors’ calculation based on IMF IFS, OECD and ECB data

Looking at the results of our estimation for selected NMS and comparing them with the average for the Club Med (Tables 3.9 and 3.10), we draw almost the same conclusion as for the high frequency data. The average stance of the NMS between 1999 and 2004 is closer to that of the Club Med countries between 1993-1995 than to their stance between 1996 and 1998. The relative magnitude of RER shocks in these two sub-samples was, respectively, 1.3 and 2.1 times higher. Since the results for the early 1990s are the same as for the high frequency data, we may conclude

58

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK

that the magnitude of both nominal shocks and real shocks is smaller for the NMS than was the case on average for the Club Med countries. In almost all cases the magnitude of individual quarterly RER variances was slightly higher than that of monthly changes19. Given that we assume unexpected quarterly RER volatility to reflect real shocks, it is clear that asymmetric shocks are still an important source of RER volatility for the NMS. Also, it is hard to say whether the reported long-run volatility decline was due to policy changes, or to common shocks hitting those countries. Compared with monthly changes, the decline was significant only for Bulgaria, Estonia, and Slovenia between the second and the third sub-sample, and for Latvia and Lithuania between the first and the second sub-sample. But even then the size of shocks in Poland and Romania was 2.5 and 2.0 times bigger (respectively) than the Club Med average of 1993-95. Turning to individual cases of selected member states, as in the case of monthly shocks it is only Greece that failed to lower unexpected RER fluctuations between the two sub-samples leading to EMU membership. The variance reduction was not statistically significant for France and Portugal either. As in the case of the NMS, the long-run volatility for the Club Med countries was slightly higher when compared with the short-run for the years 1996 to 1998, but marginally lower between 1993 and 1995. Comparing RER shocks with unexpected volatility of price differentials in the respective sub-samples, it is obvious that (with the exception of Greece between 1993 and 1995) the NER did not cushion real vulnerabilities, and that exchange rate uncertainty could be eliminated by joining the MU. 6. SUMMARY AND CONCLUSIONS In addition to reviewing the most important theoretical literature on OCA and empirical papers related to the NMS, we have attempted to assess the degree to which the NMS are ready to join the Eurozone. We found that these countries are already very open to trade with the EU, in many cases much more open than OMS. While the share of exports to the EU-15 in GDP for the Eurozone amounts to 16%, analogous indicators reach 15% for Poland, 28% for Slovenia, 36% for Estonia and Hungary and 38% for the Czech Republic. Static business cycle correlations provide a different picture. With the exception of Hungary and Slovenia, most measures of real activity co-movements point to weak or even negative correlations of shocks in the Eurozone and respective NMS. The situation is particularly problematic in the case of unemployment rates, which for most countries exhibit negative correlation with Eurozone unemployment changes. Using the nominal and RER stability criteria, and comparing them with those of the Club Med countries in the years preceding the formation of the EMU, our analysis showed that the NMS as a group resemble the Club Med countries in the early rather than mid-1990s. Two countries – Estonia and Slovenia – exhibit RER fluctuations similar to, or lower than, the Club Med countries. As for Bulgaria, the Czech Republic, Hungary, Latvia, Lithuania, Poland, Romania and Slovakia, our results suggest that RER variability still exceeds that of the Club Med countries20.

OCA CRITERIA FOR NEW MEMBER STATES

59

However, bearing in mind that the Club Med countries had to maintain their exchange rates in a +/- 15% band without devaluation for two years prior the EMU accession (and that the European Commission assessment of the countries’ eligibility to enter the EMU was even more severe and based on their avoiding depreciation of more than 2.25%), we can conclude that the average unexpected RER volatility in the NMS does not dramatically differ from that in the Club Med countries during 1993-1995, and for countries like Estonia and Slovenia it is even lower. Even though it was found that quarterly RER fluctuations slightly exceed monthly changes in almost all countries, indicating that asymmetric real shocks remain a relatively important source of RER variation, it is difficult to treat this as an argument against EMU membership. This is because the same is true for the Club Med countries, and even for France and Germany. Clearly, in the NMS, NER instability is still higher than it was in the Club Med countries before the introduction of the Euro. But the fact that NER in the Club Med countries (and in the NMS) did not suppress real volatility indicates that efficiency gains can be achieved once a MU is formed (i.e. through the elimination of exchange rate uncertainty, contagion effects, etc).

NOTES 1

Even if the exchange rate was allowed to float. Appreciate in the case of tight policy and depreciate in the case of the loose policy. 3 The high correlation between the growth rates of the Eurozone and its members is to some extent a result of the fact the aggregate figure is a weighted average of individual members’ GDP growth rates. 4 In place of averages of four annual-change indices that yielded one observation per year, the annual indices obtained for each quarter will be used to yield four observations per year. 5 Because of unavailability of quarterly data for the early 1990s, data for the Eurozone was replaced by data for Germany – the biggest member of the Eurozone. 6 The notable exception is Latvia, which has a SDR-based peg. Nevertheless, the structure of the economy is much more similar to that of other NMS and it already has tight economic linkages with the EU. Moreover, the fact that Latvia pegs to a currency basket like the SDR can mitigate the impact of exchange rate fluctuations among major international currencies. 7 The additional reason why we decided not to estimate a VAR model is that there are many studies of this kind carried out for the NMS. On the other hand, there are only a few studies in which the variance approach has been applied. 8 It should be stressed however, that before joining ERM-II, the Bank of Slovenia engineered a continuous depreciation of the exchange rate. Since its participation in ERM-II, the tolar has been trading close to its central rate of 239.640 SIT per EUR (European Commission, 2004). 9 The 2003 devaluation of the central parity of the Hungarian Forint contributed significantly to its volatility. In 2004 the Forint has been again more volatile in the context of increasing inflation, a large fiscal deficit and uncertainty related to the future policy of the National Bank of Hungary. 10 RER indexes as well as conditional variances for quarterly and monthly data were estimated with the use of the same definitions. 11 The sample includes the NMS-8 plus Bulgaria and Romania as future NMS. Henceforth, the category of NMS, if not stated otherwise, includes these two countries. 12 Gros & Hobza (2003), however, look at observed rather than unexpected exchange rate variability. 2

60

M. BLASZKIEWICZ-SCHWARTZMAN AND P. WOZNIAK

13

As a benchmark, we looked at RER variances of the Club Med countries, usually considered as members of the EU “periphery”, as well as at France and Germany, which belong to the so-called ‘core’ group. 14 The magnitude of RER volatility for Club Med countries against DEM turned to be the same, or slightly higher than the volatility of RER computed against the ECB reference rate and therefore it will not be presented here. 15 Seasonally adjusted series were obtained as deviations from the 12-month centered moving average. 16 In order to choose the appropriate number of lagged terms we applied a ‘general-to-specific’ method of estimation. The tests used were the LM test for autocorrelation, Q-statistics and Akaike and Schwarz info criteria. 17 We also performed the CUSUM of Squares Test, as this test helps assess not only parameter but also variance instability. As the results of the CUSUMSQ test were in line with those obtained by White and ARCH tests they did not change our conclusions and therefore will not be presented here. 18 Quarterly STD are presented on a monthly basis for comparative purposes. 19 This result is somewhat in contrast with that of Gros & Hobza (2003). However, they do not include Bulgaria and Romania in their analysis. When we excluded these two countries from our sample, quarterly volatility was only higher between 1999 and 2004. 20 Even if Bulgaria and Lithuania are already in the Eurozone with their Euro-denominated currency boards, their long-run RER vulnerability is greater than that of the short run, suggesting that more adjustment may be required.

REFERENCES Bratkowski, A. & Rostowski, J. (2001). Why unilateral euroization makes sense for poland and some other candidate countries? Paper presented at the Conference on ‘Polish Way to the Euro’, National Bank of Poland, Falenty. Ching, S. & Devereux, M. B. (2000). Risk sharing and the theory of optimal currency areas: a reexamination of Mundell 1973. HKIMR Working Paper, 8, Hong Kong Institute for Monetary Research. Ching, S. & Devereux, M. B. (2003). Mundell revisited: a simple approach to the costs and benefits of a single currency area. Review of International Economics, 11 (4), 674-691. Egert, B. & Kierzenkowski, R. (2003). Asymmetric fluctuation bands in ERM and ERM-II: lessons from the past and future challenges for EU acceding countries. William Davidson Institute Working Paper, 597. European Commission (2004). Convergence report 2004 – Technical annex. A Commission services working paper, SEC(2004) 1268, Commission of the European Communities, Brussels, , October 20, http://europa.eu.int/comm/economy_finance/publications/european_economy/convergencereports200 4_en.htm Patterson, B. & Amati, S. (1998). Adjustment to asymmetric shocks. Economic Affairs Series, ECON104. European Parliament, Directorate- General for Research, Working Paper Fidrmuc, J. (2001). The endogeneity of the optimum currency area criteria, intra-industry trade, and the EMU enlargement. Discussion Paper, 106, LICOS, Centre for Transition Economics, Katholieke Universiteit Leuven. Fidrmuc, J. & Schardax F. (2000). More ‘Pre-Ins’ ante portas? Euro Area Enlargement, Optimum Currency Area, and Nominal Convergence, Transition, 2, 28-47. National Bank of Austria, Vienna. Frankel, J.A. & Rose, A.K. (1998). The endogeneity of the optimum currency area criteria. Economic Journal, 108, July. Gros, D. & Hobza, A. (2003). Exchange rate variability as an OCA criterion: are the candidates ripe for the Euro? Working Papers, 23, International Centre for Economics Growth European Centre, Macroeconomic Studies. Gurjarati, D.N. (2003). Basic econometrics (4th ed.). McGraw-Hill. McKinnon, R. (1963). Optimum currency areas. American Economic Review, 53, 717-725. McKinnon, R (2000). Mundell, the Euro and optimum currency areas. Journal of Policy Modeling, 22(3), 311-324. Mundell, R. (1961). A theory of optimum currency areas. American Economic Review, 51, 657- 65.

OCA CRITERIA FOR NEW MEMBER STATES

61

Mundell, R (1973). Uncommon arguments for common currencies, [in:] Johnson, H. & Swoboda, A. (eds.). The Economics of Common Currencies. London: George Allen & Unwin Ltd., 143-173 Vaubel, R. (1976). Real exchange rate changes in the european community - the empirical evidence and its implications for european currency unification. Weltwirtschaftliches Archiv, 112, 429-470. Von Hagen, J. & Neumann, M. (1994, May). Real exchange rates within and between currency areas: how far away is EMU? The Review of Economic and Statistics, 76 (2), 236-244.

MARYLA MALISZEWSKA

CHAPTER 4

EMU ENLARGEMENT AND TRADE CREATION

1. INTRODUCTION The aim of this study is to look at the potential impact of euro adoption by the new member states (NMS) of the European Union (EU). We begin with a discussion on the theoretical foundations and empirical evidence behind the argument that adoption of a common currency (CC) increases trade flows between members of a monetary union (MU). The seminal paper by Rose (2000) who found that a CC triples trade led to a still ongoing debate on the topic and a series of new empirical studies. It seems that even though the estimated impact of a CC on trade has been slightly reduced in later studies, the evidence for a causal relationship between a MU and trade is very strong. This is also the case in early estimates of the impact of the euro. It seems therefore that if the NMS were to adopt the euro, the trade flows between them and EU-15 can be expected to increase as well. In the subsequent sections we estimate the potential increases in trade following NMS accession to the Economic and Monetary Union (EMU). First, a gravity equation similar to Rose’s original specification is estimated for the EU countries. The results of this study confirm that the euro has already contributed to greater trade flows between EMU members. Further, assuming that the same relationship between income, distance, common borders and other country characteristics on the one hand and bilateral trade on the other will hold in the future for trade between the EU-15 and NMS, we estimate the potential trade increases following the accession of the NMS to the EMU. 2. HOW CURRENCY UNIONS AFFECT TRADE BETWEEN MEMBERS Initially, academic economists were skeptical about the argument that the creation of a CC leads to greater market integration and increased trade linkages. First, in theory importers and exporters could hedge exchange rate uncertainty. Second, empirical studies found little evidence that exchange rate variability had any adverse effect on trade. However, forward and futures markets are not available for most trading partners and they entail transaction costs. The problem with the empirical argument against the impact of a CC on trade, is that it was mostly based on time series evidence, where other influences on trade could not be fully accounted for. In 63 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 63-74. © 2006 Springer. Printed in the Netherlands.

64

MARYLA MALISZEWSKA

addition the evidence was based on large industrialized countries1. When small countries were included in the analysis some effects of a CC started to show up, especially in studies of bilateral trade. The seminal contribution to the research on the impact of MU on trade was provided by Rose (2000) and subsequent papers e.g. Glick and Rose (2001). Rose (2000) included several small countries and dependencies that adopted currencies of larger countries to form a large sample of countries in MUs. He found a statistically significant negative effect of exchange rate variability on bilateral trade flows. Rose also finds that countries that share a currency, trade three times as much as otherwise similar countries with different currencies. Rose’s result provoked an ongoing debate and has raised many questions. First, the statistical link between MU and trade may not represent the impact of a CC on trade, but rather the impact of some third factor such as colonial history or others. Therefore countries with previous strong links and high trade might have decided to form a MU. It is then inappropriate to infer that forming a MU would triple trade. Second, the cross-section evidence does not allow us to evaluate the time pattern of the effect of a MU on trade. Third, the effect seems to be simply too big to be believable. Forth, Rose’s estimates came from a population of very small and poor countries. Therefore several economists doubted whether his results were relevant for larger and more developed countries. On the other hand, it is possible to defend Rose’s results on several grounds. First, regarding the endogeneity of the exchange rate regime choice, Rose has done a thorough job of controlling for common language, colonial history and political links and a large impact nevertheless remains. Second, regarding the dynamics of the effect of a MU on trade, Glick and Rose (2001) aim to answer these questions by looking at a panel of countries over the 1948-1978 period, including countries in periods when they formed part of a MU and periods when they did not. The results indicate that joining a MU almost doubles bilateral trade among its members. This suggests that roughly two thirds of the tripling effect may be reached within three decades of the adoption of a CC. Here again the analysis is focused on small and poor countries and is limited in most cases to countries that exited a MU as opposed to countries joining them. Third, regarding the large magnitude of the estimates it is important to take into account home country bias. People trade far more easily with their fellow citizens than with people from different countries. Even when one controls for the effects of distance, trade barriers, linguistic, social and cultural differences the strong tendency to trade within the country and the lack of an ability for arbitrage to keep prices in line across different spaces remain. Canadian provinces have been found to trade twenty times as much with each other than with US states (McCallum, 1995). This number falls to three times after the introduction of the NAFTA and after controlling for other factors. One of the most likely explanations of the home country bias is the existence of different currencies in different countries (Parsley and Wei, 2001). Regarding the last criticism, as to whether Rose’s result holds for larger countries, this is better understood today, as there is now more empirical evidence for developed countries. One should also note that Rose (2000) and Frankel and Rose (2002) found no significant differences between the results between small and

EMU ENLARGEMENT AND TRADE CREATION

65

very small countries. In addition, empirical studies have established that although home bias is smaller, it also operates in the case of larger countries. To the extent that different currencies explain this effect, the currency effect should also not be limited to small countries. There have also been a few studies, which have focused on larger countries. Estevadeordal, Frantz and Taylor (2002) looking at the set of developed countries and a few large developing countries find that a participation in the gold standard increased trade by 34-72% depending on specification. Also Lopez-Cordova and Meissner (2000) estimate that common participation in the gold standard increased trade by 60%. They also found that participation in a MU doubles trade. Finally, at least three years of available data since the formation of the EMU in January 1999 allows one to test the impact of a common currency on developed countries. Barr, Breedon and Miles (BBM, 2003) estimate a gravity model (see below) for the EU and EFTA countries (except for Luxembourg and Liechtenstein) over the 1978-2001 period. The authors deal with the endogeneity problem by estimating the model by instrumental variables, treating EMU membership as endogenous and using past output and price co-movements across countries as instruments. All variables are statistically significant and have the expected signs (see Table 4.1). Output and output per capita are positively related to bilateral trade flows, distance affects trade negatively, while the existence of common borders and common language enhance trade. The results of BBM (2003) indicate that the EMU 2 has already added 24% to the level of trade among members. The reduction in exchange rate volatility has also contributed to the rise of trade, but to a much smaller extent. The study is only based on the first three years since the formation of the EMU, so it is likely that the long-run effect will be much larger. Table 4.1. Trade effects of EMU Item Currency Union Exchange Rate Volatility Log(Output) Log(Output per capita) Log(Distance) Contiguity Language EU Membership

Coefficient 0.21 -0.15 1.23 0.23 -1.24 0.18 0.26 0.45

Standard error 0.04 0.01 0.01 0.04 0.02 0.02 0.02 0.02

Note: Dependent variable: log of bilateral trade. Source: BBM (2003), IV estimation, Table 3, p. 582.

Micco, Stein and Ordonez (MSO, 2003) estimate the impact of the EMU based on 1980-2001 data on bilateral trade. The authors use two data sets: a first set including 22 industrial countries and a second set limited to EU members only. They estimate the bilateral gravity trade equations introducing dummies for membership of a free trade area, the EU and the EMU. The estimated coefficient on the EMU

66

MARYLA MALISZEWSKA

dummy is always positive and statistically significant. Table 4.2 displays a range of estimates based on the developed countries’ and EU samples. Table 4.2. Trade effects of EMU EU with country pair dummies Standard Standard Standard Coefficient Coefficient Coefficient Error Error Error 0.198 0.4 0.25 0.043 0.060 0.013 0.793 0.009 0.828 0.013 0.996 0.074 0.218 0.033 0.068 0.039 0.101 0.05 0.048 0.095 0.042 0.029 0.148 0.055 -0.194 0.165 -0.073 0.053 -0.003 0.004 -0.002 0.012 -0.003 0.004 -0.495 0.032 -0.712 0.032 0.136 0.045 -0.752 0.024 -0.733 0.037 -0.012 0.008 -0.7 0.013 0.248 0.044 0.275 0.049 0.816 0.042 0.652 0.073 2541 1001 1001 0.93 0.94 0.64

Developed countries Item EMU Log(Real GDP) Log(Real GDP per capita) Free Trade Agreement EU EU Trend Landlocked Island Log(Distance) Log(Area) Contiguity Common Language No. of observations R2

EU members

Note: Dependent variable: log of bilateral trade. Source: MSO (2003), Table 1, p.328,Table B2, p.351.

In addition, in order to avoid the problem of endogeneity of the decision to form a MU, MSO (2003) run regressions with country-pair fixed effects. The previous regressions showed that countries that adopt a CC trade more than otherwise identical countries. However, previous specifications did not allow for the isolation of the impact of the EMU itself. Inclusion of country-pair dummies allows us to estimate the impact of the euro and leave out all cross-country variation. The country-pair fixed effects replace all country-specific characteristics such as distance, common language etc. and also unobservable characteristics. In this way if for any reason two countries traded a lot before the formation of the MU this effect is captured by the country-pair dummies and does not affect the estimate of the impact of the EMU. The MSO (2003) results show that the impact of the euro is much smaller (6%) when country-pair dummies are included indicating that there is a reverse causality between intensity of trade and adoption of the CC. Since the CC reduces transaction costs in trade, countries that trade a lot are likely to gain more from the adoption of the euro. MSO (2003) results indicate that indeed high trade flows were likely to be used as a criterion for membership in the EMU. Further estimations by MSO (2003) show that the value of the coefficient of the EMU dummy is increasing over time, indicating that trade among EMU members increased in anticipation of the formation of the MU. Future EMU members traded 3.6% more in 1996 than otherwise similar countries. In 1999 EMU members traded 12.4% more than otherwise similar countries. Therefore, creation of the EMU

EMU ENLARGEMENT AND TRADE CREATION

67

increased trade among its members by 8.4% between 1996 and 1999 and a further 3 3.7% between 1999-2002 . The estimates of BBM (2003) and MSO (2003) are much lower than the original estimates of Rose, which relied on data on small developing countries. The impact of the EMU may indeed be smaller than that estimated for small countries, or its effects might be greater in the long run. Bun and Klassen (2002) update gravity estimates and make dynamic projections regarding the impact of the EMU. They conclude that the euro has increased trade by 4% in the first year, and its long-run effect is estimated to reach 40%. Overall, the existing evidence suggests that formation of a MU stimulates trade between its member states. The original estimate of Rose, that a CC triples trade, seems to be a reasonable upper limit. Even controlling for the endogeneity of the decision to form a MU and looking at the sample of developed countries and at EMU post-1999, a statistically significant relationship between a CC and trade can be established. 2.1. Methodology The gravity model of trade comes from the application of the physical law of gravity to trade. Bilateral trade between any two countries depends on their market sizes measured by GDP (an analogue of mass) and the distance between them. Due to their empirical robustness, gravity models have been extensively used to explain bilateral trade between countries and to estimate the impact of preferential trade agreements. Although early applications of gravity models have been criticized for their lack of theoretical foundations, later studies showed that with special assumptions a simpler version of the gravity model can be derived from the factor proportions model (Deardorff, 1995), or from increasing returns to scale and product differentiation models or a combination of both (Evenett and Keller, 1998; Shelburne, 2000). The present study applies the original Rose’s specification to the data on EU countries. The aim is to establish a baseline against which to compare current trade between EU-15 and NMS and to formulate predictions about the likely impact of their future EMU accession. The bilateral gravity equations of trade to be estimated are as follows:

ln Tij

D 0  D1 ln YiY j  D 2 ln yi y j  D 3Cont  D 4 Lang  D 5 Dij  D 6 Area 

D 7 Landlocked  D8 EU  D 9 EMU  H ij (4.1) where T represents bilateral trade (value of exports and imports in constant 1995 dollars), Y represents GDP (in constant 1995 US dollars), y – GDP per capita (in constant 1995 US dollars), Cont – Contiguity i.e. when two countries share a border, Lang – common language dummy, D – distance between capital cities (in miles), Area – product of two geographical areas, EU, EMU – common membership in the

68

MARYLA MALISZEWSKA

EU, EMU. The EMU dummy takes a value of 1 for pairs of EMU members even before the formation of the EMU. GDP captures the economic size of countries and represents the potential for export supply and import demand. The product of countries’ GDPs is expected to affect trade positively. The product of GDPs per capita captures the notion that rich countries tend to trade more intensively than poor countries. Countries with higher level of development (as measured by per capita GDP) have more sophisticated economies and more varieties of goods on offer. If consumers value variety rich countries will be more likely to trade with similar partners. On the other hand more populous countries tend to have a higher degree of autarky and less need to trade. This hypothesis finds support in the EU data, as large countries such as France, Germany or Spain have lower shares of trade in GDP than less populous EU neighbors (see Table 4.5). Regardless of the underlying reasons in both cases the product of GDPs per capita will have a positive impact on bilateral trade. Geographical distance represents transport costs. Distance increases the costs of trade and is therefore expected to be negatively related to trade. Countries with a common border usually share some historic ties and cultural similarities, and also tend to have better knowledge of the neighboring markets. Common language reduces transaction costs. These two variables are expected to affect trade positively. Landlocked countries tend to trade less as they face higher transport costs (water transport tends to be cheaper than other ways of transport). Finally, preferential trading agreements reduce transaction costs due to lower barriers to trade and simpler border requirements, while the EU membership is also expected to enhance trade due to the existence of the Single Market. The specification applied in this study is very similar to one of the specifications of MSO (2003). I also use the same data4. However, I drop the insignificant dummy variables and I limit the estimation to EU members over the period of 1992-2002. 2.2. Results Table 4.3 presents the results of the estimation of equation 4.1. The model fits the data very well and all variables have statistically significant coefficients with the expected signs. Larger, richer countries, which share a common border or language, tend to trade more. Distant and landlocked countries tend to trade less. Naturally the results are very close to those obtained by MSO (2003) as presented in Table 4.2. The negative coefficient of the EU dummy indicates that the only non-EU countries in the sample i.e. Austria, Finland and Sweden over 19921994 period traded on average more with the remaining EU countries before their accession than EU members with similar characteristics. The effect of a CC on trade is estimated to amount to 26.5%.

EMU ENLARGEMENT AND TRADE CREATION

69

Table 4.3. Results of the estimation of the bilateral gravity equation Item EMU Log(GDP) Log(GDP per capita) Free Trade Area EU EUTrend Landlocked Log(Distance) Log(Area) Contiguity Common Language R2 No. of observations Country Pair Dummies Time Dummies

Coefficient 0.235* 0.827* 0.035

Standard error 0.044 0.014 0.035

-0.123*

0.056

-0.698* -0.779* -0.057* 0.261* 0.765* 0.94 857 No Yes

0.039 0.036 0.015 0.052 0.084

Coefficient 0.062* 0.871*

Standard error 0.013 0.047

-0.005 0.023 -0.085

0.003 0.030 0.047

0.99 857 Yes Yes

Note: * - significant at the 5% confidence level.

3. POTENTIAL TRADE FLOWS Assuming that the same relationship between trade, GDP and other explanatory variables will hold for trade between NMS and the EU-15, we can calculate their potential trade flows using the estimated coefficients presented in Table 4.3. Further, assuming that the impact of the EMU is going to be the same as for the current EMU members we can also calculate the potential impact of the EMU membership on EU15 - NMS trade. Table 4.5 displays the share of trade with the EU-15 in GDP in 2002 for the NMS-8. It shows that Estonia, the Czech Republic, Slovakia, Hungary and Slovenia are the most open economies among the NMS. On the other end, Poland’s share of trade with the EU-15 in GDP is much lower than in the other countries. We might therefore expect that Poland will be among the countries which would be likely to expand their trade with the EU by significant amounts, while the most open countries are likely to gain relatively less. Potential trade flows between the NMS and the EU-15 are obtained by taking the coefficients displayed in Table 4.3 and plugging in the values of explanatory variables for the NMS (see Table 4.4 for results). If we do not include the EMU dummy, the resulting trade potential can be interpreted as the level of trade typical for the EU members before formation of the EMU. In several NMS the level of trade with the EU-15 in 2002 was already higher than trade between old member states (OMS) would have been had they exhibited similar levels of income, location and other characteristics. However, Slovenia, Poland, Latvia and Lithuania would record increases in trade if they were to reach the levels of integration typical for the EU15.

MARYLA MALISZEWSKA

70

Table 4.4. Estimated potential trade flows between EU-15 and NMS.

Country

Czech Rep. Estonia Hungary Latvia Lithuania Poland Slovakia Slovenia

2002 Actual Trade with EU15 million USD (1) 26771 3067 23717 2139 3176 30941 8713 7311

Potential Trade with EU15 if NMS Traded as much as OMS

Potential Trade with EU15 if NMS Traded as much as EMU members

Potential increase in trade following EU membership

Potential increase in trade following EMU membership

million USD

million USD

growth in %

growth in %

(2) 18724 3145 11494 3394 4460 50470 7457 14777

(3) 23321 3979 14545 4295 5644 63866 9436 18699

(2)/(1) -40.7 -13.2 -58.9 34.4 19.0 38.2 -27.5 71.2

(3)/(1) -26.2 9.9 -48.0 70.1 50.5 74.9 -8.3 116.7

Source: own calculations.

Table 4.5. Trade with the EU-15 as a share of GDP at current prices in 2002 Country Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden United Kingdom

Trade with EU-15/GDP 22.7 54.2 18.3 17.2 14.5 15.7 10.0 36.9 11.2 44.7 35.3 22.8 15.4 19.4 11.3

AVERAGE

16.8

Country Czech Rep. Estonia Hungary Latvia Lithuania Poland Slovakia Slovenia

Trade with EU-15/GDP 38.5 47.1 36.0 25.5 23.0 16.4 36.8 33.3

26.5

Source: own calculations

This trade expansion between NMS and the EU-15, can be mainly attributed to their gaining access to the Single European Market, and the resulting reduction in real transaction costs following the elimination of border formalities, harmonization of product and safety standards and regulations, and greater similarity of the business environment. Other trade policy changes include integration into the EU customs union, freer movement of labor and incorporation of the NMS into the

EMU ENLARGEMENT AND TRADE CREATION

71

CAP. The impact of Single Market access on trade seems to be the largest of these (see Lejour, de Mooij and Nahuis, 2001; Maliszewska, 2004). The potential trade flows exceed the 2002 trade level by more than 70% in Slovenia. Lithuania, Latvia and Poland could potentially increase their trade with the EU-15 by 20-40%. The Czech Republic, Slovakia, Estonia and Hungary trade already more than an average EU-15 country given their location and income levels. Another way to look at the level of trade integration between OMS and NMS is to analyze the level of trade with the EU-15 as a share of GDP. This is a rather crude method, as trade depends on many other characteristics apart from GDP but it illustrates that NMS intensity of trade with the EU-15 is high and higher than in several OMS (see Table 4.5). Similar findings can be found in Nilsson (2000) or Gros and Gonciarz (1996). Finally, our calculations indicate that Hungary, the Czech Republic and Slovakia have already reached the level of trade integration with the EU-15 that would have been enjoyed by members of the EMU with similar characteristics. Since the increase in trade following the adoption of the euro is proportionate to the impact of EU membership the distribution of gains in trade among the NMS remains similar. The potential for trade increases between the EU-15 and Slovenia is the highest, followed by Poland, Latvia, Lithuania and Estonia. The results of our calculations do not imply that the trade of Hungary, the Czech Republic and Slovakia with the EU will drop following the introduction of the euro in these countries. On the contrary, the impact of joining the MU is expected to be positive for all countries and is determined by the size of the EMU dummy. Our calculations only indicate that these countries have already reached a level of integration with the EU typical for EMU members with similar characteristics, while other countries still trade below these levels. If the impact of a CC is indeed the same for all the economies once they adopt the euro, these three countries’ trade with the EU-15 could expand by 26%. On the other hand, the remaining four countries, which still trade below potential, could expect greater trade increases overall as they first need to take full advantage of the Single Market access and then adoption of the CC. 4. TRADE AMONG THE NMS Are the NMS trading as much with each other as the EU-15 member states do? To find out we apply the coefficients discussed above to assess the current and potential trade between the NMS. The results are displayed in Table 4.6. Again the four most open countries i.e. Estonia, Hungary, the Czech Republic and Slovakia traded in 2002 more with each other than two EU countries with similar income, geographical location and other characteristics would have traded.

MARYLA MALISZEWSKA

72

Table 4.6. Estimated potential trade flows between NMS.

Country

Czech Rep. Estonia Hungary Latvia Lithuania Poland Slovakia Slovenia

2002 Actual Trade with EU15 million USD 4470 413 2413 497 760 3664 3280 822

Potential Trade with EU15 if NMS Traded as much as OMS

Potential Trade with EU15 if NMS Traded as much as EMU members

Potential increase in trade following EU membership

Potential increase in trade following EMU membership

million USD

million USD

growth in %

growth in %

2358 323 2314 508 1046 5869 1386 1892

3113 427 3055 671 1381 7748 1830 2498

-47.2 -21.7 -4.1 2.2 37.6 60.2 -57.7 130.3

-30.4 3.3 26.6 34.9 81.7 111.4 -44.2 204.0

Source: own calculations.

Once the NMS join the EMU and reach the same level of integration as OMS, trade among them is expected to increase significantly. Slovenia would need to triple, while Poland would need to double, its trade with remaining NMS to bring it in line with the relationship between income and trade typical for the EMU members. Only Slovakia and the Czech Republic trade even more than two EMU countries with similar characteristics. This is most likely due to the fact that despite splitting in 1993 the two countries are still very integrated. The wedge between actual and potential trade flows is driven mainly by high levels of trade between the two countries and not by the level of exchange with the remaining NMS. The model in its basic specification is not able to capture this kind of historical ties between countries. This exercise, as with any gravity equation based estimates of potential trade, should be treated with caution. The assumption that the relationship between explanatory variables will hold for every country is very problematic. It is clear from the estimation of the equation with country pair fixed effects that there are other factors influencing trade which we do not take into account. As a result potential trade flows following the introduction of a CC might be greatly overestimated. 5. CONCLUSIONS The empirical evidence presented in this paper confirms the existence of a significant causal relationship between the introduction of a CC and trade expansion. The original estimate of Rose (2000), that a CC triples trade, seems to be a reasonable upper limit. Even controlling for the endogeneity of the decision to form a MU and looking at the sample of developed countries and post-1999 EMU, a statistically significant relationship between a CC and trade can be established. The early estimates for the EMU indicate that its members trade on average between 6% and 26% more than otherwise identical countries would.

EMU ENLARGEMENT AND TRADE CREATION

73

The coefficients of the gravity model have been employed to study the potential trade between OMS and NMS and between the NMS themselves. Our calculations indicate that four countries (Estonia, Hungary, the Czech Republic and Slovakia) have already reached the level of integration that would have been enjoyed by OMS if they shared similar income levels and other characteristics with the NMS. In addition all of them except Estonia record the levels of trade typical for EMU members with similar characteristics. In the remaining countries i.e. Latvia, Lithuania, Poland and Slovenia the potential trade flows exceed the 2002 trade levels by 50-70%. Although following the introduction of the euro international trade is expected to expand in all NMS in accordance with the estimates of gravity models, the levels of trade of Latvia, Lithuania and Poland are expected to rise even further. Potential trade flows between NMS following the adoption of the euro by far exceed the 2002 trade flows for all countries except the Czech Republic and Slovakia, whose result is dominated by the high historical level of bilateral trade. Slovenia’s trade flows could triple, while Polish trade flows with other NMS could double once the countries reach the level of integration typical for members of the EMU.

NOTES 1

For surveys of the literature, see Edison and Melvin (1990). 24% is exp(0.21) in the IV estimation. 3 The calculations of percentage increases in trade are based on the coefficients of EMU dummies as displayed in Table 2, p. 332. 4 I am very grateful to the authors and to Danielken Molina for making their data available to me. 2

REFERENCES Barr, D., Breedon F. & Miles D. (2003). Life on the outside: economic conditions and prospects outside euroland. Economic Policy, 517-613, October. Bun, M. & Klaassen F. (2002). Has the euro increased trade? Tinbergen Institute Discussion Paper, 02108/2. Deardoff, A.V. (1995). Determinants of Bilateral Trade: Does Gravity Work in a Neoclassical World? NBER Working Paper, 5377. Edison, H.J. & Melvin, M. (1990). The determinants and implications of the choice of exchange rate regime, in monetary policy for a volatile global economy, ed. by Haraf W.S. and Willett T.D., Washington: AEI Press. Estavadeordal, A, Frantz, B. & Taylor A. (2002). The rise and fall of world trade, 1870-1939. NBER Working Paper, 9318, November. Evenett, S.J. & Keller, W. (2000). On theories explaining the success of the gravity equation. NBER Working Paper, 6529. Frankel, J.A. & Rose A.K. (2002). An Estimate of the ffect of common currencies on trade and income. Quarterly Journal of Economics, 117(2), 437-466. Glick, R. & Rose, A.K. (2001). Does a currency union affect trade? The time series evidence. NBER Working Paper, 8396. Gros, D. & Gonciarz, A. (1996). A note on the trade potential of Central and Eastern Europe. European Journal of Political Economy, 12, 709-721.

74

MARYLA MALISZEWSKA

Lejour, A.M, de Mooj, R.A. & Nahuis R. (2001). EU enlargement: economic implications for countries and industries. CESifo Working Paper, 585, CESifo, Munich, Germany, October. Lopez-Cordova, E. & Meissner C. (2000). Exchange-rate regimes and international trade: evidence from the classical gold era. University of California, Berkeley, Center for International and Development Economics Research (CIDER) Working Paper, C00/118, November. Maliszewska, M. (2004). EU enlargement: benefits of the single market expansion for current and new member states. CASE Studies and Analysis, 273. McCallum, J. (1995). National borders matter: Canada-US regional trade patterns. American Economic Review, 85 ( 3), 615-623. Micco, A., Stein E. & Ordonez, G. (2003). The currency union effect on trade: early evidence from EMU. Economic Policy, 315-366, October. Nilsson, L. (2000). Trade integration and the EU economic membership criteria. European Journal of Political Economy, 16, 807-827. Parsley, D.C. & Wei, S. (2001). Explaining the border effect: the role of exchange rate variability, shipping costs, and gography. Journal of International Economics, 55 (1), 87-106, October. Rose, A. K. (2000). One money, one market: estimating the effect of common currencies on trade. Economic Policy: A European Forum, 30, 7-33, April. Shelburne, R.C. (2000). Intra-industry trade, the gravity model and similarity in endowments and country size. Papers and Proceedings of the International Trade and Finance Association, San Diego State University.

ARTUR RADZIWILL AND MATEUSZ WALEWSKI

CHAPTER 5

FUTURE EMU MEMBERSHIP AND WAGE FLEXIBILITY

1. INTRODUCTION Future membership of the Economic and Monetary Union (EMU) is obligatory for the EU new member states (NMS). However, the policy discussion in the NMS recently has been somehow short-sighted in its exclusive focus on the timetable of meeting the Maastricht nominal and fiscal convergence criteria. We believe more attention should be paid to factors that will decide whether membership in the monetary union (MU) proves successful in the medium term and its implications for the design of structural policies. In this chapter, we recall that it is the structure of the labor market (LM) that is of particular importance for minimizing the potential costs of EMU membership. Wage flexibility and free international migration is essential for adjustment to asymmetrical shocks to occur; when exchange rates are irrevocably fixed, monetary policy is driven by union-wide considerations and scope for fiscal policy response is limited. We attempt to evaluate the risks of increased size and volatility of unemployment in the NMS after accession to the EMU by studying past LM adjustment mechanisms and, in particular, the role played by exchange rate movements and inflationary surprises. We conclude that risks from the EMU might be lower than commonly believed, but governments have a significant role to play by actively promoting the reform of LM institutions. The remainder of this chapter is constructed in the following way: section 5.2 discusses the importance of LM flexibility within the MU, section 5.3 presents empirical evidence on patterns of adjustments in selected NMS, and section 5.4 discusses our conclusions. 2. THE THEORY OF OPTIMAL CURRENCY AREAS AND BEYOND According to Mundell (1961) and his theory of Optimal Currency Areas (OCA), nominal wage flexibility is the perfect substitute for nominal exchange rate (NER) flexibility1. If, in response to rising unemployment, local currency wages and, consequently, the price of domestic products falls, then the competitiveness of the economy is improved. Therefore, the speed at which nominal wages fall determines

75 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 75-90. © 2006 Springer. Printed in the Netherlands.

76

ARTUR RADZIWILL AND MATEUSZ WALEWSKI

the unemployment cost of asymmetric shocks in MU. The implications of OCA theory for empirical work seem to be straightforward. Countries with more flexible nominal wages are better prepared to enter MU. Countries with nominal wage rigidity should seriously consider whether the benefits of EMU participation outweigh the possible costs resulting from LM maladjustments. Unfortunately, the conclusion of Jackman & Savouri (1998) is that ‘it seems unreasonable to think that degree of wage flexibility that characterizes European labor markets can substitute for exchange rate flexibility’.

Accordingly, in view of observed rigidities, many researchers have questioned the net benefits and even the viability of the EMU (Feldstein, 1997). However, we believe that any research strategy that focuses exclusively on econometric evaluation of wage rigidities will be unsatisfactory due to a number of issues that we discuss below. 2.1. The role of productivity growth The first and simplest consideration against an exclusive focus on nominal wage elasticity is that it ignores the role of productivity growth. The potential costs of MU are related to decreased competitiveness due to nominal downward wage stickiness in the absence of exchange rate depreciation (of greater importance for more open economies), or the rigidity of real wages due to low and stable inflation (of greater importance for more closed economies). But what matters for hiring and firing decisions are domestic and international unit labor costs (DULC and IULC), which also depend on the level of productivity. They are defined respectively as: DULC

IULC

W E˜P*

W P

Q L

(5.1)

Q L

R ˜ DULC

(5.2)

where: W is nominal wage level, Q is real output, L is employment, E is NER, R is real exchange rate (RER), P and P* are domestic and foreign price levels, respectively. When productivity grows rapidly (what is expected to happen as part of real convergence), both measures of labor costs fall despite rigid wages. Accordingly, in response to a negative shock, rising productivity may by itself (without the need for wage reduction) lead to a relatively quick reduction of labor costs, permitting employment to be restored. A rarely formulated implication emerges: faster growing NMS are likely to be ceteris paribus more successful members of the EMU, not because of faster nominal adjustment, but thanks to high productivity growth.

FUTURE EMU MEMBERSHIP AND WAGE FLEXIBILITY

77

2.2. Doubts about benefits of flexible exchange rates and national monetary policy Another problem relates to an implicit assumption about the positive adjustment role of exchange rate flexibility and independent monetary policy. Firstly, it is assumed that NER movements and independent monetary policy actions aim to help to equilibrate LM. Here many authors express serious doubts. Vinals (1996) has observed that ‘… recent experience suggests that the usefulness of the nominal exchange rate as a tool for macroeconomic adjustment within the European Union is very questionable in a world of free capital movements, where foreign exchange markets are often subject to self-fulfilling speculative crises which take the exchange rate away for prolonged periods from where fundamentals suggest it should be.’

Ochel's (1997) observation on the role of financial shocks is particularly important here: ‘Asymmetrical financial shocks can be better dealt with in a monetary union than in a system with adjustable exchange rates and are thus not an additional source of unemployment.’

Even more importantly, Bayoumi & Eichengreen (1994) underline the role of policy-induced shocks: ‘Policy makers may systematically misuse policy rather than employ it to facilitate adjustment, [...] if domestic policy itself is the source of the disturbances, monetary unification with a group of countries less susceptible to such pressures may imply a welfare improvement.’

The invalidity of the assumption about the usefulness of flexible exchange rates and independent monetary policy might easily reverse the implications of the theory of the role of LM flexibility: if exchange rate movements and independent monetary policy are the sources of shocks, rather than useful adjustment mechanisms, the country with the least flexible wages has the largest interest in joining an area of stable inflation2. Therefore, the potential costs of abandoning flexible exchange rates and independent monetary policy for LM cannot be assessed without an analysis of their effectiveness. Secondly, in order to believe that currency depreciation or an expansionary monetary policy can indeed reduce unit labor costs (ULC) when unemployment is above its equilibrium rate and nominal wages are rigid, it is necessary to assume, respectively, that: (i) nominal depreciation might lead to improved competitiveness or a reduction in real euro-wages W/EP*, (ii) wages are not rigid in real terms so that inflation can reduce W/P. While the positive role of an inflationary policy requires at least temporary money illusion, effective depreciation requires money illusion or a low correlation between exchange rate movements and domestic price levels, i.e. flexible RER P/EP*. It follows that high wage rigidity does not need to be an argument against participation in a MU. If real wages, W/P, are rigid in the sense that nominal wage dynamics closely follow inflation, monetary regime is mostly irrelevant for LM outcomes. The same is true for smaller countries for which there is high passthrough from exchange rates to prices (rigid P/EP*).

78

ARTUR RADZIWILL AND MATEUSZ WALEWSKI ‘For small economies in which imports are so important a component of the cost of living that a reduction in real wages through nominal depreciation is rapidly feed through into domestic prices’ (Obstfeld, 1985).

Indeed, from McKinnon (1963) onwards, one of the main criteria for the choice between fixed and flexible exchange rates has been the degree of openness in an economy, characterized by relative importance of traded to non-traded goods. There is some evidence suggesting that nominal depreciation can be useful in improving competitiveness. First, real wage and price levels seem to fluctuate more among countries with independent currencies than in regions within a currency area (De Grauwe & Vanhaverbeke, 1993; von Hagen & Neumann, 1994)3. There have also been a number of recent episodes when devaluation provided a boost to competitiveness and helped reduce unemployment. The most often cited example is the UK withdrawal from the EMS in 1992: according to most authors, devaluation and monetary expansion contributed to the subsequent decade of economic recovery without visible inflationary consequences (Jackman & Savouri, 1998). Another well-known episode proves that devaluations can be ‘a great success’ even in a relatively small economy: devaluation in Belgium in 1982 had a strong and lasting positive impact on current account deficits and employment (De Grauwe, 2000). We are curious whether we can actually deduce recent similar episodes in NMS – their presence would be an argument in favor of increasing LM flexibility after their acession to the EMU. 2.3. The Lucas Critique and the prospects for labor market reform The last and most fundamental research problem is connected with the Lucas Critique: estimates derived under one policy regime have very limited value in predicting developments under another regime. When monetary policy is independent and the exchange rate fluctuates, an improvement in competitiveness or reduction in real wages does not necessarily require nominal flexibility. Observed nominal rigidity is likely to be high, which may tempt conclusions about the undesirability of MU. However, a change in the regime might in itself bring about a change in the LM flexibility. In particular, workers who were able to accept real wage cuts in a high inflation environment might potentially accept nominal wage cuts in a fully credible stable price environment. In a related argument, Calmfors (1998) believes that, given the constraints on active usage of the exchange rate and monetary policies, incentives for LM reform would increase once the EMU is in place. If this widely held hypothesis is true, any measure of current nominal rigidity is of limited relevance for the evaluation of the likely costs of EMU participation. 3. EMPIRICAL RESULTS Our investigation is based on quarterly data compiled by the IMF in its International Financial Statistics (IFS) database for six countries that joined EU in 2004: the Czech Republic, Hungary, Latvia, Lithuania, Poland and Slovakia (see Annex). Unfortunately, adequate data was not available for Estonia and Slovenia. The dataset

FUTURE EMU MEMBERSHIP AND WAGE FLEXIBILITY

79

actually used was further updated and completed with statistics from national statistical offices and central banks and covers the period between 1994 and 2002. From the discussion above, it is clear that an econometric estimation of historical wage rigidity is unlikely to give robust insights about the potential costs of membership of the EMU unless the role of the exchange rate, price and productivity movements are analyzed at the same time. Hence, in line with the discussion above, we define overall LM elasticity as the response of two described measures of ULC to changes in unemployment. Adjustments might take place through nominal wages and prices, which determine real wage dynamics (W/P), through improvements in productivity, as well as through deprecation of RER (P/EP*)4. We focus on changes rather than levels of LM aggregates as it is very difficult to identify structural and cyclical elements of unemployment in the NMS. Our preliminary econometric investigation suggested that ULC reacted to changes rather than levels of unemployment5. We are especially interested in adjustment patterns in times of rising unemployment. Obviously, it is impossible to eliminate the Lucas Critique problem completely: based on the current situation we cannot judge well how structural aspects of LM will change after EMU accession. Our strategy is, however, at least in this respect superior to mechanistic econometric investigation of wage flexibility, as we explicitly consider alternative adjustment channels. As an additional benefit we can compare adjustment mechanisms in the NMS that still conduct an independent monetary policy and those with currency board or quasi-currency board regimes. 3.1. Country episodes We analyze the period beginning in 1994, when GDP growth rates were generally positive and correlated across countries, with the exception of the Czech Republic. This common pattern was interrupted by one shock: the Russian crisis of 1998. The importance of this event is clear when we observe the unemployment dynamics. Average unemployment rates fell moderately from above 9% in 1994 to around 8% in 1998, but afterwards increased markedly to 12%. Also, until 1998 unemployment rates in most countries seemed to be converging, but then started to diverge. The spread between the highest and lowest unemployment rates in our sample was 13 percentage points in 1994, 3.5 in 1998 and 12 in 2002. So, we will focus on the Russian crisis as well as other episodes of economic meltdown in individual countries. Crisis situations provide the best opportunity to analyze reactions to adverse and asymmetric shocks. Before conducting a cross-country econometric investigation, we study different adjustment mechanisms in each country in more detail. 3.1.1. The Czech Republic The Czech Republic experienced a long period of GDP slump in the years 19961999. This slump resulted in a rapid increase in unemployment from around 4% in mid-1996 to a maximum of about 10% at the beginning of 2000. When economic growth returned, unemployment remained at this higher level. During the initial

80

ARTUR RADZIWILL AND MATEUSZ WALEWSKI

period of unemployment growth between the second half of 1996 and mid-1997, DULC continued to grow and did not help to ease the problem. Only for a short period, in 1997-98, did a sharp decline in real wages prevent larger increases in unemployment. Then DULC stabilized. Real wages did not seem to adjust to changes in unemployment, with the long run dynamics remaining more or less in line with productivity changes. Even the short lasting decrease in real wages in the years 1997-1998 had been preceded by a similar negative shock to productivity. The results of simple econometric estimation confirm that real wages are not negatively correlated with unemployment in the Czech Republic. The lack of adjustment in real wage dynamics in the Czech Republic after the initial deep slump in 1997 seems to be strictly related to nominal wage rigidity in this country. After nominal growth in 1996-1997, wages kept rising at a relatively stable rate and did not react to the pronounced unemployment increase that occurred in 1999. We can also see that the real wage decrease following the currency crisis of 1997 was more the effect of the inflation surprise than of nominal wage adjustment. Without the inflation surprise Czech workers would not have experienced a fall in real wages even during the hardest period of crisis and falling productivity. These observations imply that nominal wages in the Czech Republic are relatively sticky. As real wages grew, on average faster than productivity, and were not responsive to unemployment changes, DULC failed to adjust. The relative rigidity of DULC and their upward trend predicts long term problems in the Czech LM, whether or not this country joins the EMU. Only at during times of crisis did monetary policy prove useful in pushing unit labor cost downwards. NER movements did not seem supportive of LM adjustments either: since 2001 CZK/EUR exchange rate has appreciated despite continuously rising unemployment. 3.1.2. Hungary The unemployment rate in Hungary fell continuously from 11.5% in 1994 to 5.6% in 2001. Given such a favorable trend, which was uninterrupted even during the economic slowdown of 1994-1995, downward adjustments in DULC were not necessary and DULC remained stable through most of the period until 2001. This happened as real wages were rising in line with productivity growth. Characteristically, moderation of real wage growth between 1998 and 2001 was not the result of inflationary surprise. On the contrary, slower nominal wage dynamics were necessary because of the disinflation process. However, there are two episodes when monetary policy played important role in shaping real wage dynamics: the GDP growth slumps in 1995 and the last quarter of 2002. In the first case, high inflation combined with a decline in nominal wage dynamics (a fall from more then 25% at the end of 1994 to around 15% in mid1995) led to deep cuts in real wages and DULC, and only part of this reduction was reversed by high nominal wage increases in subsequent years. It is difficult to escape the conclusion that inflation surprise and nominal wage moderation in 1995/1996 contributed to a subsequent unemployment rate decline. On the other hand, rising real wages in 2002 were due to the absence of wage growth adjustment to fast

FUTURE EMU MEMBERSHIP AND WAGE FLEXIBILITY

81

disinflation. As a result, DULC have markedly increased, which probably contributed to increases in unemployment. Nominal wage growth acceleration at the same time as falling inflation was partly due to a statutory increase in minimum wages. This might suggest that it was a one-time effect; on the other hand, the minimum wage increase could lead to a permanent reduction in nominal wage flexibility. More importantly, the increase of DULC in 2002 also illustrates the perils of volatile inflation; after two years of fluctuating at around 10%, inflation fell rapidly to below 5% in 2002 and this change was not reflected in nominal wage setting. To sum up, Hungary seems to have rather inflexible nominal wages. The option of inflating wages proved useful during the time of national emergency in 1995. It seems that participation in the EMU at that time (had EMU existed) could have led to less favorable domestic unit labor cost and unemployment dynamics. At less turbulent times, the loss of monetary independence should not have a significant impact on LM adjustment. In particular, conservative but stable monetary policy would prevent the problems of adjustment to rapidly falling inflation that characterized 2002. Also, the strong NER appreciation at the time of rising unemployment suggests the limited usefulness of this adjustment tool. 3.1.3. Latvia Latvia is one of the two countries in our sample operating without monetary and exchange rate policy freedom. This country has a fixed exchange rate to SDR and its monetary base is permanently backed by net foreign assets at more than 100%, which leaves very limited scope for an independent monetary policy. LVL/EUR exchange rate changes cannot play the adjustment role suggested by the OCA theory, as they are caused exclusively by changes in cross rates of major world currencies. Nevertheless, during the period analyzed, the unemployment rate in Latvia seems to have been much more stable then in other countries analyzed, with the exception of Hungary. During the GDP slump in 1995-1996, the unemployment level did not change at all, presumably partly as a result of the deep DULC adjustment that took place during this time. Unemployment started to decline before the Russian crisis of 1998 in an environment of very high economic growth and this decline was accompanied by a sharp increase in real wages. Unfortunately, the Russian crisis of 1998 resulted in a sharp GDP slow down and unemployment also rose to more then 10% at the beginning of 1999. Even this rise, however, has been much smaller then in other countries, such as Slovakia, Poland or even the Czech Republic. It seems to be the result of the sharp adjustment of DULC that took place as a consequence of unemployment growth. In later years, in spite of an improving LM situation, DULC have never regained their positive dynamics, facilitating the recovery process. The negative DULC dynamics since 1998 are both the effect of moderate real wage growth between 1998-2001 and very high productivity dynamics throughout the entire period. Productivity was rising even during the actual Russian crisis. Although nominal wage dynamics fell twice in Latvia, leading to a real wage adjustment, this does not appear to be the main factor that determined the relatively

82

ARTUR RADZIWILL AND MATEUSZ WALEWSKI

good and stable LM situation. The determining factors were wage growth moderation combined with favorite productivity dynamics, which caused continuous DULC fall. This situation is different from what we have observed, especially in Slovakia and the Czech Republic and also to some extent in Poland, where inflation surprises led to the most serious real wage adjustments. The above characteristic of the Latvian LM, if preserved in the future, would continue to be a big advantage in the case of an adverse asymmetric shock after joining the Eurozone. 3.1.4. Lithuania Lithuania is a country with a currency board regime, which implies that it cannot conduct an independent monetary policy: the NER cannot restore competitiveness and inflationary surprise cannot be used to reduce DULC. Due to Lithuania’s close trade links with Russia and other CIS countries, the 1998 financial crisis in Russia sent the economy into a deep recession. Unemployment, which was at a relatively low level (6-8%), increased markedly in 1998 until reaching its peak of 13% at end 2000, after which it began to fall gradually and has continued to do so ever since. DULC response was not immediate but was impressively strong. During the last quarters before the crisis, real wages were growing much faster than productivity and real wages continued to grow moderately in 1999. However, the rapid increase in unemployment finally forced a very strong adjustment in 2000. As a result, real wages plunged and given the very low level of inflation, adjustment occurred though reductions in nominal wages. Actually, Lithuania is the only country in which an absolute fall in nominal wages, not only adjustment of its dynamics, took place. Since that time, real wages have grown more slowly then productivity, leading to constant DULC reduction. The example of Lithuania shows that deep adjustment in real wages is possible under MU if nominal wages are flexible enough. It is, however, difficult to judge whether this flexibility is a inherent feature of the Lithuanian LM or if the currency board forced downward nominal wage adjustment. The one-year lag that the actual adjustment came with is evidence for the latter being true. 3.1.5. Poland Before 1998 Poland experienced high economic growth with unemployment initially stabilized and then falling. During that time long run real wage growth followed the rise in productivity, keeping the average DULC unchanged. It seems that wage moderation facilitated the fall in unemployment rate during this period. Since 1998 unemployment began to rise very rapidly and kept rising until the very end of the period analyzed (2002). In response, DULC initially decreased sharply and thereafter kept falling as impressive improvements in productivity took place and were not reflected in higher real wage dynamics. On the other hand, the decrease in real wages in the second half of 2000 was mainly the result of inflation surprise. However, the impact of high productivity growth rates on ULC prevails and permits the expectation of rather low costs of EMU membership. This is even more so as Poland is another country where exchange rate flexibility brought strong NER appreciation in times of increasing unemployment.

FUTURE EMU MEMBERSHIP AND WAGE FLEXIBILITY

83

3.1.6. Slovakia The unemployment rate in Slovakia fell from 1995 onwards, bottomed out at around 11-12% between 1996 and 1998 and rose rapidly to 19% at the end of 1999. Smallscale fluctuations reduced the unemployment rate in 2001 and increased it in 2002. The unemployment dynamics were particularly well reflected in DULC. They were falling precisely at times of increasing unemployment. From our reading of the evidence, the role played by the independent monetary policy in lowering DULC in response to negative shocks was higher than in any other country in our sample. When, in 1999 the fallout of the Russian crisis coincided with the initiation of rapid structural reforms to produce a marked slowdown in GDP growth rates and a rapid increase in unemployment, real wages were the main adjustment mechanism. But the real wage cuts that took place in 1999 had much to do with the increased inflation, while nominal wage growth remained stable throughout the period analyzed. This nominal wage inflexibility also played a role in 2002 when the disinflation process led to an increase in both real wages and DULC. Slovakia was therefore a country in which the inflationary impulse managed to push down real wages in reaction to a rapidly increasing unemployment rate. On the other hand, nominal wages were not responsive to disinflation. Finally, improvements in productivity were not an important channel of adjustment to negative shocks. These results suggest that participation in the EMU might prove costly. However, it is also possible that nominal rigidity is due precisely to experience of an active monetary policy and would be removed when the country joins the EMU. However, this possibility could be questioned, given the lack of nominal wage adjustment to the disinflation in 2001–2002. However, a more credible monetary policy might change this maladjustment pattern. 3.2. Regression results The adjustment mechanisms analyzed in the previous subsection proved to be remarkably heterogeneous across countries. This observation is confirmed by the results of the regression explaining annual changes in the DULC presented in Table 5.1. Slovakia, Lithuania and Poland have the domestic unit LM costs that are the most responsive to unemployment changes; however, we know that adjustments took place through very different channels in these three countries. Flexibility in Slovakia depended predominantly on inflation surprise and in Poland mainly on productivity growth accelerations. Lithuania adjusted through nominal wages. Conversely, the LM in the Czech Republic, Hungary and Latvia seem to have been remarkably inflexible. Why did Hungary and Latvia exhibit such positive unemployment dynamics? On our interpretation, this was due to falling average labor costs across the cycle. Consistent real wage growth moderation below productivity dynamics proves to be very good insurance against adverse shocks. Lithuania, which generated, on average, increasing DULC, needed very deep cuts in real wages after the Russian crisis; Latvia did not. It is worth noting that the Czech Republic exhibits not only rigid but also generally increasing ULC. It is somehow surprising that Poland seems to have

ARTUR RADZIWILL AND MATEUSZ WALEWSKI

84

one of the most favorable outcomes, combining high flexibility with negative average ULC growth. If these results are correct, then high unemployment levels in Poland are due to factors which are more fundamental (e.g. qualification and regional mismatches) than a lack of labor cost flexibility. Table 5.1. Explaining the dynamics of DULC Country Czech Rep. Slovakia Hungary Poland Latvia Lithuania

ULC growth rate at stable unemployment 0.011 0.013 -0.007 -0.019 -0.032 0.010

Elasticity of DULC to unemployment growth* 0.010 -0.255* -0.011 -0.120* -0.024 -0.147*

* coefficient significant at 10% confidence level Fixed effect panel estimation, quarterly 1994:1 2002:4, adjusted R-Squared= 0.162974 Source: Own calculations based on IMF-IFS data.

Table 5.2. The correlation between nominal and real exchange rate depreciation* Country Czech Rep. Slovakia Hungary Poland Latvia Lithuania

Immediate 0.81 0.86 0.88 0.57 0.45 0.19

After 1 quarter 0.47 0.54 0.73 0.33 0.24 -0.04

After 3 quarters -0.29 -0.28 0.36 -0.08 -0.13 -0.42

Note: * quarterly data since 1994 (1995 for Baltic States) Source: Own calculations based on IMF-IFS data.

Table 5.2 shows that while changes in NER are strongly contemporaneously correlated with RER in most countries, after only three quarters the entire effect is largely eliminated and indeed reversed. Therefore, while it is possible to suspect that EUR exchange rate movements provide some short-term relief, they are unlikely to play any major role in medium-term adjustments. 4. CONCLUSIONS Summing up, our analysis indicates that, with the exception of Lithuania, nominal wages are not flexible in NMS, which is usually regarded as strong evidence against participation in the EMU. However, the degree of wage rigidity varies significantly among the countries analyzed. Fast productivity growth creates an environment in which ULC can adjust despite nominal stickiness, through moderation of real wage dynamics. This structural feature would remain crucial after accession to the EMU. For example, productivity in all NMS analyzed grew in 2002 between 2.7% and 4.7% annually. This implies that even at a stable inflation rate of 2% and full

FUTURE EMU MEMBERSHIP AND WAGE FLEXIBILITY

85

downward nominal wage stickiness, DULC can fall 4.7-6.7% annually. And we note that in recent years, with the notable exception of adjustment following the Russian crisis, none of the countries that had an independent monetary policy, recorded reductions of such magnitude in ULC. Moreover, the deeper adjustments following the Russian crisis were achieved mainly by additional improvements in productivity that were not reflected in real wages. Inflationary surprises played an important role in adjustment to major adverse shocks, triggering real wage adjustments that were not fully reversed afterwards. However, our analysis indicates that in normal times the discretionary monetary policy was more often a shock generator than a shock absorber. It is also quite possible that changes in the monetary regime would result in more nominal wage flexibility, as suggested by the example of Lithuania. NER movements played some role in adjustment after the Russian crisis. At all other times, in spite of their short term impact on RER, they did not seem to be correlated with the LM situation. It is therefore our impression that OCA theory does not shed much light on the possible unemployment consequences of EMU membership. This is especially so, since shocks of such depth and degree of asymmetry relative to current EMU members as the Russian crisis, are rather unlikely to be repeated, even though large fluctuation in the EUR/USD exchange rate may imply severe adjustment problems for some NMS. Finally, the experience of countries with the most favorable unemployment dynamics suggests that ULC falling across the cycle provides the best insurance against adverse economic shocks. There is no unanimous evidence that real wage growth lower on average than productivity growth would be harder to generate inside rather than outside MU. Moreover, if bargaining is increasingly taking place at the European level, this will be generally favorable for countries with better than average productivity dynamics, such as NMS, since low real wage dynamics might be imported from slow growing mature economies. This optimistic conclusion is obviously reversed if levels rather than dynamics of wages in existing EMU members start to influence wage bargainers in NMS. In our analysis of actual adjustment patterns in selected NMS, we remain focused on our main research question: the potential impact of EMU accession on LM performance. However, another interesting result of this research is the large diversity in LM adjustment mechanisms, even in this relatively homogeneous group of countries.

NOTES 1

High labor mobility is another substitute for flexible exchange rates. Indeed Mundell (1961) defines OCA as an area within which factors of production are mobile but outside which they are immobile. We do not study patterns of migration in this paper, but previous research suggests that ‘it is unlikely that greater mobility of labor, either within or between Member States, can ever become a major instrument of adjustment within the EU’ (Patterson & Amati, 1998). Conclusions pointing in the same directions for NMS can be found in Kupiszewski (2002) and Firdmuc (2002). 2 This point was brought to our attention by Jacek Rostowski.

86

ARTUR RADZIWILL AND MATEUSZ WALEWSKI

3

There are, however, different factors that might contribute to this result: notably that currency areas historically corresponded to countries, which through institutional setup and redistributive policies tended to compress regional wage and price differentials. 4 We study the variation in EUR exchange rates instead of effective exchange rates. The reason is that it is precisely the size and adjustment role of previous variations against the common currency that is crucial for the potential cost of joining the EMU. 5 These and other auxiliary econometric results are available upon request.

REFERENCES Bayoumi, T. & Eichengreen, B. (1994). One money or many? Analyzing the prospects for monetary unification in various parts of the world. Princeton Studies in International Economics, 76, International Economics Section, Department of Economics, Princeton University. Calmfors, L. (1998). Unemployment, labour-market reform and monetary union. Seminar Papers, 639, Stockholm University, Institute for International Economic Studies. De Grauwe, P. (2000). Economics of monetary union. Oxford: Oxford University Press De Grauwe, P. & Vanhaverbeke, W. (1993). Is Europe an optimum currency area? Evidence from regional data. [in:] Masson, P. R. & Taylor, M. P. (eds.). Policy Issues in the Operation of Currency Unions. Cambridge: Cambridge University Press. Feldstein, M. (1997). EMU and international conflict. Foreign Affairs, November/December. Fidrmuc, J. (2002). Migration and regional adjustment to asymmetric shocks in transition economies. William Davidson Institute Working Papers Series, 441, February. Jackman, R. & Savouri S. (1998). EU labour markets and monetary union. [in:] Cooper, A. (ed.). Economic Outlook. Oxford Economic Forecasting and the London Business School. Kupiszewski, M. (2002). How trustworthy are forecasts of migration between Poland and the European Union? Journal of Ethnic and Migration Studies, 28 (4) , October. McKinnon, R. (1963). Optimum currency areas. American Economic Review, 53, 717-725. Mundell, R. (1961). A Theory of optimum currency areas. American Economic Review, 51, 657- 65. Obstfeld, M. (1985). Optimum currency areas and the misalignment problem. Brookings Papers on Economic Activity. Ochel, W. (1997). European Economic and Monetary Union and Employment, Ifo Institut. mimeo. Patterson, B. & Amati, S. (1998). Adjustment to asymmetric shocks. Economic Affairs Series, ECON104. European Parliament, Directorate-General for Research, Working Paper. Vinals, J. (1996). European monetary integration: A narrow or a wide EMU? European Economic Review, 40, 1103-1109. Von Hagen, J. & Neumann, M. (1994). Real exchange rates within and between currency areas: how far away is EMU? The Review of Economic and Statistics, 76 (2), 236-244, May.

FUTURE EMU MEMBERSHIP AND WAGE FLEXIBILITY

87

APPENDIX 5.1. SELECTED ECONOMIC DEVELOPMENTS RELATED TO LM ADJUSTMENTS Source for all Figures: Own calculations based on IMF-IFS data. 0.15

0.10

0.05

0.00

-0.05

-0.10 96

97

98

99

00

01

Czech Republic Slovakia Poland

02

Hungary Latvia Lithuania

Figure 5.1. GDP real growth rates 20

15

10

5

0 94

95

96

97

98

99

Czec h Republ i c S l ovaki a Hungary

00

01

02

P ol and Latvi a Li thuani a

Figure 5.2. Unemployment rates

ARTUR RADZIWILL AND MATEUSZ WALEWSKI

88 0.3

0.2

0.1

0.0

-0.1 -0.2

-0.3 96

97

98

99

01

00

02

Poland Latvia Lithuania

Czech Republic Slovakia Hungary

Figure 5.3. DULC Growth rates 0.3

0.2

0.1

0.0

-0.1

96

97

98

99

Czech Republic Slovakia Hungary

00

01 Poland Latvia Lithuania

Figure 5.4. Productivity growth rates

02

FUTURE EMU MEMBERSHIP AND WAGE FLEXIBILITY 0.3

0.2

0.1

0.0

-0.1

-0.2 96

97

98

99

02

01

00

Poland Latvia Lithuania

Czech Republic Slovakia Hungary

Figure 5.5. Real wage growth rates 0.3

0.2

0.1

0.0

-0.1

-0.2 1997

1998

1999

2000

Czech Republic Slovakia Hungary

2001

2002

Poland Latvia Lithuania

Figure 5.6. Nominal wage growth rates

89

ARTUR RADZIWILL AND MATEUSZ WALEWSKI

90

0.2

0.1

0.0

-0.1

-0.2

-0.3

95

96

97

99

98

00

01

Czech Republic Slovakia Hungary

Poland Latvia Lithuania

Figure 5.7. Real depreciation rates (against the Euro) 0.20

0.15

0.10

0.05

0.00 1997

1998

1999

2000

Czech Republic Slovakia Hungary

Figure 5.8. Inflation rates

2001

Poland Latvia Lithuania

2002

02

MATEUSZ SZCZUREK

CHAPTER 6

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

1. INTRODUCTION The new member states (NMS) of the EU will have to decide when and how to join the Economic and Monetary Union (EMU). The question ‘how’ includes the desired path to fulfillment of the Maastricht criteria, the strategy for fixing the exchange rate, and the actual final EUR/local currency exchange rate. This chapter considers the second of the above three issues. It assesses the relative merits of various options of the ERM-II setup with regards to the fulfillment of the Maastricht criteria, as well as the impact of exchange rate regimes leading to the ERM-II. It also presents theoretical considerations on the likely consequences and exchange rate dynamics of the regime switch from the original, pre-EU accession currency band to that of the ERM-II, and to the king of all pegs, the monetary union (MU). The chapter begins with the theoretical arguments behind the choice of different variants of the ERM-II in terms of ease of meeting the Maastricht criteria (section 6.2). Section 6.3 evaluates how the choice of pre-ERM-II regime can influence nominal convergence. Section 6.4 looks at the exchange rate dynamics of the regime switch leading into the ERM-II. Section 6.5 concerns the US Dollar legacy, and the speed of adjustment of the local financial markets’ focus from the USD to the common currency. Section 6.6 concludes. 2. THE MAASTRICHT CRITERIA AND ERM-II OPTIONS 2.1. Inflation In theory, during the latter stage of transition, a floating exchange rate makes the fulfillment of the inflation criterion easier. Fast growing NMS are prone to the Harrod-Balassa-Samuelson (HBS) effect, which leads, under a fixed exchange rate, to inflation in non-tradable goods which is not compensated for by falling prices of tradables1.

91 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 91-111. © 2006 Springer. Printed in the Netherlands.

92

MATEUSZ SZCZUREK

A second channel that works in a similar fashion is the decreasing cost of capital, which is likely to be observed in the intermediate and latter stages of transition. Assuming the tradable sector is more capital intensive, a given fall in capital costs will increase the marginal return on labor to a greater degree in that sector. This in turn implies growth of wages and prices in the non-tradable sector, and consequently the real exchange rate will also have to appreciate (Buiter & Grafe, 2002). A third channel, mentioned by Rostowski (Chapter 1 of this volume) is the demand channel – higher growth of demand for non-tradables in fast growing economies puts pressure on wages and prices in that sector, and on the real exchange rate to appreciate (provided productivity in non-tradables does not grow too fast). The impact of all these effects to equalize relative prices in the NMS and old member states (OMS) is 3.5-4% per year, according to Pelkmanns, Gros & Nunez Ferrer (2000), and this is the potential gain in terms of CPI inflation from having the exchange rate able to appreciate in nominal terms. From this point of view, the classic ERM-II, with a ±15% band is best, as it leaves enough room for nominal appreciation, provided that the initial market rate is close to the parity rate. The strict ‘Solbes version’ with a ±2.25% band may not leave enough room to fulfill the exchange rate criterion, leading to excessive pressures on domestic prices. Bearing in mind the possibility of revaluing the parity without violating the exchange rate stability criterion, any non-currency board version of the ERM-II is likely to be better at tackling the problem of growth-related real exchange rate appreciation. The cost of altering the currency board rate makes such changes less likely. Because the expected EMU parity usually remains in the centre of the ERM-II bands (Ireland and Greece did have their exchange rate revalued, but other catchingup countries did not), rational economic agents should set their prices and wage demands to take into account the expected final year depreciation of the local currency (if it has previously appreciated below the ERM-II parity). This could limit or completely eliminate the beneficial effect of a semi-floating exchange rate on inflation. The Drachma and Escudo started to depreciate back towards the central parity some one year and eight months ahead of EMU entry, while in Spain the process started three years ahead of the country’s joining the EMU. If ‘convergence check’ comes some eight months ahead of accession to the EMU, and exchange rate pass-through is significant for periods less than one year, the inflation effect of the ±15% ERM-II could actually work in the opposite direction to the one described in the previous paragraphs. Also on the negative side, non-fixed exchange rate regimes may promote shortterm volatility in the case of global (or regional) market disturbances. Habib (2002), and Csermely & Vonnak (2002) show that world emerging markets’ contagion does significantly influence the exchange rates of the Polish and Czech currencies (and the Hungarian Forint, with a currency band regime). This means that a major world market shock could prompt depreciation, leading to problems with the CPI inflation criterion. In the final stages of the EMU accession the argument for emerging market problems is already less valid, but is more valid for converging markets. As the date of the likely EMU entry approaches, the Hungarian Forint will be less prone to, let say, Latin American or Russian turbulence, but more susceptible to, say, fiscal disturbances in the Czech Republic. This is exactly what happened by the end of

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

93

2003, when concerns over Polish fiscal policy prompted another wave of Hungarian Forint selling and another bout of bond spread divergence. Thus, such events could, theoretically, cause some inflation-related problems which are unjustified by local policies or fundamentals. A major capital outflow, however, is unlikely to cause inflationary problems for the exchange rate regime, provided it fails to topple the exchange rate arrangement. Firstly, for fixed exchange rate regimes and narrow bands, the effect of capital outflow can be contractionary. A negative inflationary impact for wide bands will only appear under extremely precise timing. It would have to happen about one year ahead of the Maastricht check, and even then, the inflationary impact could quickly be reversed if belief in the parity returns (as it should). Another exchange rate regime-related inflation issue comes into play if the exchange rate band becomes binding. Capital inflow (appearing for example as a result of insufficient savings in economies that are expecting faster growth in the post-EMU years) being absorbed by the central bank with higher base money can become a problem either for inflation or for the budget, through the costs of sterilization. Provided the parity is not set much weaker than the market rate, the problem will be much more pronounced for the ±2.25% ERM variant, or for currency boards (although the latter do not engage in sterilization). The ±15% band is able to emulate a float more easily in this respect2. Disentangling the impact of the ERM-II arrangement on inflation is thus quite difficult. The standard relative price adjustment argument for the floating exchange rates cannot easily be translated into the superiority of the ±15% ERM-II band over the narrower arrangements, currency boards, and unilateral Euroization. This it due to the fact that nominal appreciation can be expected to be reversed in the final stages of EMU accession. One problem which a wide band is likely to have on a smaller scale is monetary expansion occurring when the exchange rate hits the bottom of the band. Here, pegs, currency boards, and narrow bands are inferior, provided that the initial parity of the ±15% band is not set too weak relative to the market rate. 2.2. Interest rates Success in fulfilling the interest rate criterion depends largely on the credibility of the EMU accession, i.e., on the ability (or willingness) to fulfill all other criteria. If the markets believe the country will be considered fit to join the EMU, they will make sure that bond yields adjust roughly to EMU levels. The effect of public debt levels, bond supply, and fiscal stance on the Bunds, other EMU member bonds’ spreads, is very limited (within 30bp, judging by current EMU members’ bond spreads vs. Bunds). What is more, while inflation inertia is a significant factor to bear in mind while preparing to join the EMU, bond yields can be extremely quick to adjust – almost jumping into the new, low-yield equilibrium at the moment of entry into the ERM-II. This makes the choice of both pre-ERM-II exchange rate regime and the actual variant of the ERM-II regime somewhat irrelevant for the ability to fulfill the interest rate criterion at the time of the EMU entry decision.

94

MATEUSZ SZCZUREK

Theory suggests that a floating exchange rate allows for nominal appreciation through the HBS effect, savings-investment imbalance created by expectations of higher growth, and stimulation from opening and deepening of the NMS financial markets. Due to uncovered interest parity (UIP), this would suggest lower bond yields than in a more rigid exchange rate set-up, where real appreciation translates into higher inflation pressures. But, as shown above, inflation and bond yields in the ±15% ERM-II band may not benefit from the nominal appreciation pressures. This is because these are likely to be reversed back towards parity in the final 1-2 years of the ERM-II in any case. A 5% nominal depreciation expected in one year would then translate into 120bp higher spread on 5-year bonds. Because there is a non-zero probability that the parity would be revalued, the negative impact of the final year depreciation expectations could be somewhat limited. Moreover, the wide-band ERM-II, which is different from the float from the point of view of the mean expected value, can be almost equivalent to the float on the risk front. This could additionally boost the bond spreads. A short-term currency premium, which is bound to be higher than in the case of a credible peg or currency board, may have an important effect particularly on the short-end of the yield curve. Should therefore the currency board ERM-II result in unconditionally lower yields than a narrow-band ERM-II, which, in turn, has lower yields than the ±15% arrangement? Not necessarily: capital inflow-related inflation problems in the more rigid versions of ERM-II could make the whole process of accession doubtful, putting additional spread burden on the economy. The overall impact of the exchange rate regime on bond yields is therefore unclear, and depends on where the market exchange rate is relative to the likely parity (for both versions of the bandERM-II, especially the +/-15% variant), and on the inflation performance in the year prior to the Maastricht criteria check. 2.3. Exchange rate stability and the EMU parity There are three issues related to the exchange rate in the run-up to the EMU. The first is the exchange rate stability criterion itself, and the impact of the choice of exchange rate regime on the ability to fulfill it. The second is the desired exchange rate. The third issue is the switch from the ERM-II regime towards the final fixing. Several points need stressing before we engage in a comparison of exchange rate regimes. First, the ERM-II is not symmetrical. Revaluations of the band are permissible, meaning that a successful speculative attack against the Euro does not have to breach the criterion. Such an attack could succeed if defense of the stronger edge of the band were to threaten the inflation criterion significantly. Second, the ERM-II is not the exchange rate regime of a single country. It is a matter of common concern; the NMS will not be alone in setting the parity. It is easier to start with an evaluation of the exchange rate stability criterion for the pegged regimes. The literature on the subject is vast, starting with papers following the Krugman classic (1979), in which a macroeconomic policy incompatible with the peg causes gradual depletion of reserves. A fixed exchange regime can last only until foreign exchange reserves reach a certain critical level,

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

95

when rationally-thinking speculators attack and buy all remaining stock of reserves as soon as the ‘shadow’ price – the price that would prevail without the central bank fixing the exchange rate – reaches the official rate. The regime turns smoothly to a float (the exchange rate does not jump; only the level of reserves does). Krugman & Rotemberg (1992) provide a model of imperfectly credible exchange rate bands. It joins two strands of the exchange rate literature: target zones models and currency crisis models3. In the following s is the log of exchange rate, m is the log of money supply, D is nominal domestic credit, R is nominal foreign exchange reserves, a change in v is a shock to the money demand following random walk with a drift:

s

§ ws · m  v  JE¨ ¸ , wv © wt ¹

Pwt  Vwz , m = ln(D + R)

(6.1)

Under a freely floating exchange rate, expected depreciation is equal to the drift in the money demand. Holding money supply constant, the general solution of the model, is4:

s

m  v  JP  AeDv  BeD v ,

(6.2)

where A and B are free parameters, and

D

 JP  J  P   2JV 

JV 

! 0; D 2

 JP  J  P   2JV 

JV 

0

(6.3)

With the exchange rate fully floating, money supply is truly constant and cannot be expected to change, thus the expectations component is simply equal to the expected change in v – the drift. In such a case both A and B are equal to zero. If, however, the central bank is expected to spend reserves defending some target exchange rate, then rational speculators expect money supply to fall as soon as the exchange rate reaches the edge of the band. The size of A and B depends on the amount that the government is willing to spend defending the edges, the size of the reserves in the case of A, and the amount of reserves the central bank will buy in case of an ‘attack’ on the stronger side of the band (which determines B). Thus A can be determined by knowing the level of the foreign exchange reserves, and no predictable discreet jumps in the exchange rate can happen. After a speculative attack when the reserves r are exhausted, the money supply m falls to the log money demand d, and the regime is floating. Therefore the shadow exchange rate is:

š = d + v +Ȗµ

(6.4)

MATEUSZ SZCZUREK

96

1

R  e D  1 ), as D before an attack occurs when the shadow exchange rate is equal to the regime exchange rate. This is the only rate that eliminates the possibility of capital gains for the speculators. The attack happens when v reaches the critical level v’ (for which š is equal to the targeted level of the exchange rate smax): For a small level of reserves (reserve/domestic credit ratio

d  v'JP

m  v'JP  AeDv

(6.5)

The free parameter A in Equation 6.5, which ensures that the exchange rate does not jump in a discrete fashion, is equal to:  e smax d JP D  r . Figure 6.1 shows the exchange-rate-money demand shock loci for the free-float (dashed line) and zero-reserves (post attack) float (solid straight line, parallel to the free float). The latter can be considered a shadow exchange rate. It becomes the actual exchange rate at a point where it crosses the curved line (target regime). While it is obvious that the zero-reserves curve shows a stronger exchange rate than the (no-attack) free-float (total money supply is lower by the amount of reserves, so the exchange rate must be relatively stronger after the attack), the fact that the target regime curve is below the free float is more exciting. It shows that despite the inability of the government to defend the target rate (as soon as v reaches v’ the regime collapses and the exchange rate starts to follow the zero-reserves curve), the exchange rate is supported by the sheer willingness of the authorities to spend reserves defending the target (Krugman’s ‘honeymoon effect’). The kinked thick curve (X-Y-Z) is the actual exchange rate-money demand schedule. s 10

Z

s 10

š

8

8 smax

Y

6

6

4 X

4 2

2 v’

2

4

6

8

10

v

2

4

6

8

10

Figure 6.1. Target exchange rate with limited (left) and high (right) reserves

The situation looks different when the reserves to domestic credit ratio is bigger. For large reserves, the regime curve reaches its maximum to the left of the postattack locus. A speculative attack then does not take place at all, and the target can be maintained with very small interventions. This type of situation is shown in the

v

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

97

right panel of Figure 6.1. A is set to make the regime locus tangent to the exchange rate target (s cannot be expected to grow above the smax without the attack). Each intervention, however, shifts the regime locus to the right, as the reserves get smaller 1

R e D  1 ), i.e. when the D maximum of the regime locus is at the (smax,v’) point, a speculative attack occurs that consumes all the remaining reserves and the regime turns into a free float. The important point from this analysis is that an attack cannot occur with sufficiently large reserves. If the initial reserves are much larger than domestic credit, the reserve loss is zero when the fundamentals are good enough (v is low). As v gets worse, the reserves start to dribble out (along the horizontal part of the bold curve in the right panel of Figure 6.1). As they keep worsening, an attack occurs at some stage (the drift in the money demand shock term ensures that in the original model),eliminating all the remaining reserves. We would argue that this first-generation model with monetized budget deficits is most unlikely to be of use for the NMS. Since governments cannot force central banks to monetize deficits, such a risk for any ERM-II member is non-existent (currency-boards are by definition protected from such behavior). Also, as mentioned above, it is the strengthening trend in real fundamentals that the countries will have to cope with. Thus for fixed exchange rates the reserves will not dribble out, but dribble (or rush) in. Still, the framework used for currency band models can be used to accommodate almost any drivers of the exchange rate. Second generation models5 allow governments to optimize when deciding on an exchange rate regime. The loss function usually includes the exchange rate and some variable dependent on both the actual depreciation and prior public expectations of depreciation. In the two models presented in Obstfeld (1994), the variable is the level of taxation (dependent on nominal interest rates, and thus on public expectations of nominal depreciation), or unemployment (dependent on agents’ wage setting decisions, and thus nominal depreciation). This class of models appears much more likely to trouble the ERM-II members, not least because it seems to describe well what happened within the ERMI in 1992. What would cause such a forced departure from the peg? Very high unemployment could cause the markets to believe that it would be cheaper politically to give up exchange rate stability (and prompt EMU entry) for a depreciation. For a politically feeble government, substituting difficult structural reforms for the quick fix of major depreciation could appear tempting, especially if they were not expected to be in office to reap the political glory of EMU entry. New models have also appeared that explore asymmetric information issues. Governments' implied bail-out promise distorts investment decisions and leads to contingent government liabilities either becoming actual fiscal spending, or killing the banking sector through a wave of insolvencies (see for example Corsetti, Pesenti & Roubini, 1999) or liquidity crises (building on bank run literature, and maturity mismatch in capital account-surplus countries, see for example Rodrik & Velasco, 1999). Such problems could trouble the NMS regardless of the exchange rate regime, especially if the process of excessive private foreign borrowing (on the and smaller. When the reserves reach the ‘small’ limit (

98

MATEUSZ SZCZUREK

investors’ presumption of insurance) and failed investments begin significantly before entry into the ERM-II. In this case the cumulative crisis could be large enough to topple not only the currency peg, but also any kind of ERM-II band. Similar logic can be applied to liquidity issues – once foreign debt builds up, a sudden financing stop could cause a severe crisis due to the costly liquidation of investments. The models described above can usually be applied to both fixed and band exchange rate arrangements. The NMS are unlikely to engage in reckless money printing, which could lead to a first-generation type crisis, and fundamentals are likely to put pressure on the stronger edge of the band (additionally, the currency board is protected by high reserves and automatic monetary tightening in the case of speculative pressure). The case of policymakers willing to devalue (and abandon the ERM-II) to jumpstart an ailing economy at the expense of early EMU adoption is more problematic. Incentives created by the European Commission, the ECB and the OMS may influence the loss function of the NMS’ policymakers. Outright opposition to NMS’ entry into the EMU will increase the temptation to sacrifice EMU entry in favor of stimulating economies. The mere hint of such a thing happening could increase market pressure and prompt the failure of the ERM-II. On the other hand, the costs of abandoning a currency board (especially a long-lasting one) could be much higher than for the band, so rational speculators would be less willing to pay a hefty interest rate spread to speculate against such a currency. 2.4. The switch from the ERM-II to the EMU The switch from the ERM-II to the currency union is likely to be close to a nonstochastic process switching environment. Thus a model describing the behavior of the financial markets faced with a pre-announced currency peg in a clearly defined future should be of some use. Obstfeld & Stockman (1985) ascertained that the announcement of a known peg on a known date leads to an immediate jump towards the new parity (relative to the original free-floating exchange rate path), and a smooth adjustment towards the preannounced peg. Ichikawa, Miller & Sutherland (1990) applied such a regime change to the credible exchange-rate-band model. Their findings are consistent with those of Obsfeld & Stockman (1985): the announcement of a fully credible peg (exchange rate band of zero width) results in an immediate jump of the exchange rate towards the future peg (the jump is small if the regime change is distant enough). The exchange rate then depends less and less on the fundamentals (money supply and velocity of circulation), and more on the exchange rate level set by the authorities. At time 0 (EMU entry), the exchange rate becomes totally insensitive to the underlying fundamentals. Djajiü (1989) provides another model, based on the standard monetary model, in which the exchange rate path ahead of the fixing at a known date depends on public expectations concerning the length of the fixing period (it is expected to last forever

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

99

in case of EMU entry) and the desired ratio of foreign exchange reserves to domestic money after the fixing. These models are applicable to a situation with a fully credible announcement of both the level and the time of the regime switch, and this should describe the environment of the final 6 months of ERM-II membership well. While the timing of entry should be known relatively far in advance, the exchange rate would come slightly later.

s

F

C’ s v C F Figure 6.2. Exchange rate and fundamentals on the way from ERM-II to EMU

The inherent asymmetry of the final peg expectations may have some impact on the market exchange rate while in the ERM-II. The Krugman’s S-curve cannot remain symmetrical if fixing the exchange rate weaker than the parity cannot be expected6. Thus, as the date of EMU entry approaches, the S-curve should become much flatter on the weak side of the parity (the exchange rate would be less and less sensitive to fundamentals because of appreciation expectations in the credible ERMEMU switch). The change of the shape of the S-curve on the stronger side of the parity would depend on the political bargaining process, but given the experience of previous EMU entry cases, the benchmark case of ERM-II parity becoming the final rate would suggest the curve gets flatter on that side as well (see Figure 6.2). 2.5. The Budget deficit and public debt There are several channels through which the choice of the exchange rate regime influences the fiscal position of the government. First, a non-rigid exchange rate, in the context of the NMS, can stimulate nominal appreciation. This in turn hits the profits of (and the tax intake from) the non-tradable sector. The total effect of lost taxes depends on productivity growth in the sector. Similarly, non-tradable sector inflation (and thus the nominal PIT and CIT in the sector) can be contained by nominal appreciation (relative prices in the tradable-non-tradable sector can change without wage and price increases). In the short-run this also reduces VAT revenues7.

100

MATEUSZ SZCZUREK

For countries that have managed to remove most forms of automatic inflation indexation (those countries that have recently been through a period of sub 5% CPI inflation), a period of higher inflation does facilitate fiscal consolidation. Nominal downward rigidities are prevalent in labor market regulations in most NMS (see Chapter 5). Almost all taxes depend on nominal wages and prices, while some kinds of spending are not indexed to the inflation rate. The biggest gains, both in terms of output and fiscal balance, can be achieved by surprise depreciation and inflation. This is out of the question within the wide band ERM-II, but a NMS can devalue and set a more rigid exchange rate in order to get some inflation. Getting the balance right (improving the fiscal balance while not overdoing it on the inflation side, which could breach another Maastricht criterion) is tricky. The exchange rateinflation and exchange rate-output pass-through factors differ among the countries, so the size of the effects would vary. Even though the inflation effect of the currency board or a peg need not be much larger than for a currency band (because of the relatively small likelihood of nominal appreciation becoming permanent through parity revaluation), the average exchange rate during the ERM-II period is likely to be stronger for the band than for the peg8. The second channel through which the exchange rate regime influences the budget deficit is debt servicing. As argued above, a floating exchange rate can have a two-fold effect. Expectations of nominal appreciation (as well as lower inflation) should keep the yields and debt service costs down. Despite the non-zero probability of a revaluation, such expectations would be severely limited by the benchmark expectations of a return to the parity, as was the case for most of the current EMU members. On the other hand, short-term exchange rate risk can still influence a large part of the yield curve. A third indirect influence of the exchange rate regime on budget deficit is related to the ‘straightjacket’ of the fixed exchange rates. Advanced NMS cannot resort to deficit monetization. However, the perils of slack fiscal policy under a fixed exchange rate or under a currency board are well known, and are also well understood by the policymakers in the NMS. Not a single fixed exchange rate NMS (including Bulgaria) had a budget deficit above 3% of GDP in 2002, while all others (apart from Slovenia and Romania) had. Even if hard economic arguments do not fully support the case for a higher budget deficit under the floating exchange rate regime (a possibility of nominal appreciation for the +/-15% band and the related lower debt financing costs, but higher short-term currency risk widening the spreads), the absence of the straightjacket of a fixed exchange rate does seem to make the non-pegging NMS less strict about their budgets. The public debt impact of the ERM-II regime choice can be due to three factors. The first is the budget deficit, which is likely to be smaller under more rigid exchange rate regimes, and especially under a currency board. The second factor is inflation, and its influence on nominal GDP, the denominator of the public debt ratio. Again, lower inflation, which should be associated with a (properly aligned) wide ERM-II band, puts such countries in a disadvantageous position.

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

101

The third is the exchange rate itself. Nominal appreciation has an immediate public debt reducing impact as the local currency value of Euro-denominated debt falls. Such a process, in turn, favors wide-band ERM-II members the most. The overall influence of the exchange rate regime depends on the currency composition of public debt. For example, nominal appreciation would certainly be negative for the Czech Republic, which has hardly any foreign public debt. For Poland, where 33% of the country’s debt is in foreign currency, the short-term impact of appreciation would be positive. For countries with a significant share of foreigncurrency debt, and short ERM-II membership, the band is likely to be positive for meeting the public debt Maastricht criteria. Longer ERM-II membership favors pegging as a way to keep public debt levels in check. Indeed, the data seem to support it – floating exchange rate countries (as are almost all of the NMS) have low public debt (36% of GDP on average), but that is still ten percentage points more than in the case of the pegging countries, including Bulgaria with its debt of 65% of GDP (the average public debt of the other three was a mere 14.7% of GDP). 3. THE CHOICE OF THE PRE-ERM-II REGIME Can a specific regime choice before the ERM-II make it easier to meet the Maastricht criteria later on, within the ERM-II? It seems to matter for just three Maastricht criteria: inflation, the budget deficit, and public debt. Interest rates and exchange rate stability do not exhibit sufficient inertia to influence the ability to fulfill these criteria one year ahead of the EMU. 3.1. Inflation For advanced NMS a floating exchange rate should help in bringing inflation close to the target easier in the long run, which, given inflation inertia, should help in meeting the Maastricht criterion. Many countries seem unable to use any means of fighting very high inflation other than a currency peg. The cases of Romania and Bulgaria are good examples. Romania, with a (dirty) floating exchange rate, had 22.5% inflation in 2002, the highest of all the NMS. Bulgaria, on the other hand, had the highest inflation rate of all pegging accession economies, at 5.8% compared to 2019.5% in March 19979. Even so, the performance of the NMS-2004 group suggests it is unlikely that the peg will be the only inflation-credible exchange rate regime by the time of EU entry for Romania and Bulgaria. Relative price adjustment will thus be more of a problem than post-transformation price stabilization. Adjustable but managed exchange rate regimes demonstrated less favorable performance as an inflation-fighting tool because of the competing goal of supporting external competitiveness. Lower inflation-fighting credibility caused, on average, over four percentage points higher inflation in managed, but not fixed, exchange rate regimes. In the later stages it is likely that the temptation to devalue, which is presumably higher in non-fixed managed regime countries than in currency

102

MATEUSZ SZCZUREK

board regimes, may boost inflationary expectations even more in such countries prior to the adoption of the ERM-II. What is the actual performance of the NMS so far? It matches the indicated inflation performance relatively well. If we exclude Bulgaria and Romania from the calculations, the average floating NMS’ CPI inflation in 2002 was 2.34, while for the pegging/currency board countries it was 1.92% YoY. Hungary and Slovenia with their intermediate regimes fared the worst of the advanced NMS with 6.4% inflation. Including Bulgaria and Romania meant the currency board and pegged exchange rate countries losing their lead to the floaters by 0.5 percentage points. Managed, but not pegged countries fared worse still with 11.8% average inflation in 2002. 3.2. The Budget deficit and public debt Due to nominal rigidities it may be easier to reform public finances with higher inflation. This effect, however, is likely to be limited for post-transition economies, with indexation mechanisms still not having been completely removed. The speed of debt accumulation depends as well on bond yields. Nominal appreciation expectations favor the floaters again. However, they suffer from higher short-term foreign exchange volatility. This factor may be even more important than the underlying long-term strengthening trend, which should be seen in NMS with low-inflation floating exchange rate regimes. For longer bonds, the ones that are actually instrumental for the convergence play (investments based on the presumption that the interest rates will converge into the core-EMU levels), another type of exchange risk is also crucial. This is the risk of an up-front depreciation at the gate to the ERM-II, which is almost the last chance of an officially sanctioned devaluation. While all countries face the temptation to devalue, differences still exist. It is hard to imagine that current parities of the Eurolinked currency boards would be altered, which puts Estonia, Lithuania and Bulgaria in slightly more comfortable positions. Policymakers of other exchange rate regime economies could attempt to use a temporary weakness of the currency to establish the ERM-II parity. The facts on bond yields are mixed. It appears that long-term bond yields have not converged more in floating exchange rate countries than in the currency board ones. Even though the Czech Republic represents the lowest spread (where yields were, at one stage, trading through the German equivalents), the yields and inflation performance of the other countries with floating exchange rate regimes are no better than those of the peggers. Probably the most important fiscal difference between currency regimes in the run-up to the ERM-II is the recognized value of tight fiscal policy in very rigid exchange rate arrangements. Slack fiscal policy in such countries increases problems with inflation and external competitiveness. It can also cast doubt on the regime altogether. This forces the currency board and rigid peg countries to reform their finances, which subsequently makes it easier to meet the Maastricht criteria on debt and deficit.

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

103

4. THE REGIME SWITCH Does the exchange rate regime matter for the dynamics of setting the exchange rate parity in the various versions of the ERM-II? 4.1. From float and managed float As has already been shown above, models describing the transition from a float to a credible, predefined currency peg in known time (in the case of a fixed-exchange rate version of the ERM-II) point to a jump in the exchange rate towards the peg, and a slow loss of the relationship between monetary fundamentals and the exchange rate. This is because the expectations component of the exchange rate determination becomes more important as the time of the switch approaches (which, in turn, arises from the monetary credibility of the peg). Ichikawa, Miller & Sutherland (1990) show the dynamics of the conversion from a float to a perfectly credible currency band (similar to a change from, say, a floating Czech Koruna exchange rate to the ERM-II). The analytical solution quickly becomes intractable, but the basic idea is similar to the one proved for transition from a float to a fix. The currency moves from a linear relationship between fundamentals and exchange rate to a familiar S-shaped curve, only slightly bent at first, and in time 0 becomes fully contained within the exchange rate bands. The size of the jump depends on how far in advance the change is announced. If policymakers fail to inform the markets well in advance, they can be faced with a significant jump in the exchange rate (but not in the interest rate). The announced peg or band may not be credible for two reasons. Either the parity points to an overly strong local currency, or the parity is set too weak in relation to fundamentals or market conditions. Too strong an exchange rate (relative to the floating rate) requires major monetary tightening to be credible, leading to growth problems. The overall fiscal consequences would depend on the debt levels – the primary deficit would be expected to rise (due to growth-related erosion of the tax base), while debt service costs would fall as a result of yields dropping, thanks to expectations of nominal appreciation. Too weak a parity may prove to be unsustainable either because of the threat to the inflation targets or because of the quasi-fiscal costs related to the necessary money market sterilization operations10. In order to avoid monetary policy shocks while targeting a specific exchange rate, policymakers of a floating NMS could resort to state contingency of the regime switch. According to Smith & Smith (1990), after World War I British policymakers wanted to (and let everyone know about it): i) fix the Pound to gold; ii) have the parity stronger than the post-war exchange rate, preferably at the pre-war level; iii) leave some time for the exchange rate to adjust to the desired level, at which point the fixing would occur. The situation of a floating NMS is not much different. In particular, the markets: i) expect the government/central bank will fix the national currency to the Euro; ii) often believe that the government would like to see the local currency weaker than the average market rate; iii) see that the floating exchange rate regime creates the possibility to do this.

MATEUSZ SZCZUREK

104

What are the implications of such a regime-switching environment? Smith & Smith (1990) claim that in the case of Britain it resulted in a longer return to the gold standard and a weaker Pound during the adjustment. They got this surprising result by assuming that the Pound had an exogenous tendency to strengthen, which was expected to be counteracted on as soon as the Pound reached the desired parity level.

s s*

F' C F

F'

s

F

F s

C v

v

v

F Figure 6.3. State (left) and time and state (right) dependent regime switches

The left panel of Figure 6.3 illustrates this in a set-up similar to Krugman & Rotemberg’s Figure 6.1 F’F’ is a free floating relationship between fundamentals and exchange rate, which assumes expectations of a long-run nominal appreciation (e.g. due to relative price adjustments). FF is a free floating relationship without such expectations – it points to a weaker exchange rate than F’F’ (the difference depends on the expectations parameter, and the drift in fundamentals). If the authorities promise to fix the exchange rate as soon as it reaches the level s*, rational speculators must expect an end to the fundamental appreciation trend at that point. Therefore the fixing point must lie on the FF line (from that point onwards no appreciation is expected). Thus, the intermediate regime must be between the two curves, as the expectations of appreciation slowly give way to the reality of fixing. The conclusion of the argument is that the government’s known wish to make use of the appreciation trend to fix the parity at a stronger level leads to a weaker exchange rate than otherwise might be expected in the run-up to the fixing. As Miller & Sutherland (1992) point out, there are two problems with the application of such a paradox to Britain’s return to gold in 1925. Firstly, the claim that the switch was not at all time-contingent is unjustified. Similarly, it can be claimed that the NMS would not be willing to wait forever for the exchange rate to reach the proper level. The timing of the switch could occur earlier if conditions proved right, but still no later than a specific date (for example, one determined by the political cycle formula: election date minus 2 years to allow for EMU entry

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

105

before elections). Assuming money velocity does not follow any trend, the switch would look like that shown in the right panel of Figure 6.3. The free float exchange rate (s)-monetary shock (v) locus is shown as the 45 degree line FF. The desired exchange rate is s-bar. The credible announcement of the fixing within some time frame (but not necessarily at a certain date) shifts the exchange rate-monetary shock locus up for v
106

MATEUSZ SZCZUREK

key is to show investors that C in Figure 6.3 is very high and that devaluation is neither feasible nor desirable. Means to achieve this could include convincing the public that other EU MS will not tolerate competitive devaluation. While in case of unilateral Euroization EU institutions have very limited influence on the parity rate, both EMU and ERM-II entry requires agreement of all the EMU member states, according to Article 123 (5) of the Treaty. Does this mean that an up-front competitive devaluation is impossible? Not entirely; politics rules all such decisions, and there is a possibility of key EMU MS agreeing in exchange for fishing rights or CAP reform for example. The second possibility is to convince the public that devaluation/depreciation is not desired, for whatever reason. While this is not easy given the model, the authorities can at least stop talking about the need for a weaker currency. It is simply costly. The third possibility is pegging the currency quickly and permanently. The final option is indexing the debt to a foreign currency. This way, devaluation would be counterproductive for reducing the public debt. Also useful are binding debt level limits expressed as a percentage of GDP. The Maastricht limit of 60%, as well as the equivalent Polish constitutional limit may make devaluation impossible, as it would influence not only the numerator, but also the denominator (nominal GDP level expressed in Euros). Thus, with only 10% foreign debt, but an overall debt level of 60% of GDP, any depreciation would require a costly fiscal adjustment program, even if inflation was to make up for the loss in nominal GDP. Theoretical models of a perfectly credible regime change announcement point to a gradual shift of the currency towards a new parity, with fundamentals losing their importance as the critical day approaches. The situation with an endogenous parity level or an endogenous switch time is more complicated, but leads to interesting conclusions. First, the markets have reason to expect the government to use the endof-game argument to devalue and wipe out its domestic currency-denominated debt. This could lead to an increase in bond yields and a weakening of the market exchange rate prior to the parity rate announcement. Second, in the unlikely case of the government setting the parity stronger than market rate or of an exogenous trend in exchange rate (caused for example by relative price adjustments), the convergence of the nominal rate to the desired level would be slowed down. 4.2. From a fixed exchange rate For NMS facing a fundamental real appreciation trend, pegged exchange rates and currency boards can normally last forever. The typical first-generation crisis model dynamics shows growth in domestic money offset by falls in reserves: money supply must be held constant in Flood & Garber (1984), as PPP and uncovered interest parity ensures prices and interest rates do not change, and neither does money demand while the exchange rate stays fixed. One way to describe the NMS situation is to lift the PPP condition by adding non-tradables’ prices, which exogenously drift upwards. This ensures that nominal money demand grows together with overall

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

107

price level, which, if domestic credit remains constant, boosts reserves in each period of the peg11. With the announcement of a float, money supply becomes constant, and changes in the non-tradable price level should translate into expected nominal appreciation and lower interest rates. The change depends on the rate of non-tradables’ inflation, their weight in the consumer price index, and the interest rate elasticity of money demand. A sudden switch to a float would then appreciate the currency immediately. The announcement of the end of the peg some time ahead of the event would ensure no exchange rate jump. It would boost the money supply (through higher growth of foreign exchange reserves) to the point that the peg exchange rate is also a free float solution. This growth in money supply would ensure interest rates temporarily drop ahead of the end of the peg to maintain money market equilibrium amid prices responding to the (still pegged) exchange rate and a constant trend in non-tradable prices. The transition to the currency band would depend on the parity chosen, and the results are shown in the two panels of Figure 6.4. A surprise switch into the currency band is equivalent to a vertical jump from a point on the FF curve (where money demand increases are offset by reserve growth) to the S-shaped curve CC, based on free float F’F’ (where money demand increases are not offset by money supply changes), and expectations of monetary expansion/contraction at the edges of the band. For the currency band parity set at the peg level, the step appreciation for immediate regime change would be smaller than for the switch to a float. The currency could appreciate even if the parity is set weaker than the peg. It would appreciate more than in the floating case if the parity is set much stronger than the peg (unlikely in the NMS case).

s

F

s

F'

F'

C

C'

C

s

s

v

v C

C

C'

F

F

F F'

F'

Figure 6.4. Surprise (left) and pre-announced (right) switches into a currency band

As in the floating case, prior announcement of the currency band (likely to happen in the ERM-II) would imply an inflow of reserves, and a temporary interest rate fall for a wide range of parity options, and capital outflow for an extremely weak parity.

MATEUSZ SZCZUREK

108

A special case would occur if the peg were to be outside the currency band. Setting the stronger edge of the band weaker than the peg would imply the commitment to print as much local currency as was needed to get the currency within the band. Such a possibility, along with a peg still in place until the moment of the introduction of the band, would require an anticipated jump in the exchange rate, providing infinite profits for speculators. Such a situation should never occur in competitive and open financial markets. If the peg is too weak relative to the preannounced band, the result would be an infinite inflow of capital and a forced appreciation of the local currency. 5. FORGETTING THE DOLLAR Large foreign exchange volatility on G-3 markets highlights a problem for some of the NMS. EMU entry could actually increase the exchange rate volatility faced by some economic agents. This could happen if the EUR/USD exchange rate is more volatile than the prior USD/local currency exchange rate. Below, we propose a method of measuring attachment of the local exchange rate to key world currencies. As a result, we calculate the share of the Euro, SEUR, in an optimal basket which minimizes currency risk (while financing capital inflow or investing abroad)12.

S EUR {

VarUSD  CovEUR,USD VarUSD  VarEUR  2 ˜ CovEUR,USD

(6.6)

SEUR is the proportion of the Euro which the investor (or local firm financing itself from abroad) should choose to minimize its risk to EUR/USD rates. Rom an ia

L at via

100%

100%

90%

90%

80%

80%

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10% 0%

0% A pr-99

Apr-00

A pr-01

A pr-02

A pr-03

A pr-04

Apr-99

A pr-00

A pr-01 Apr-02

A pr-03

A pr-04

Figure 6.5. Euro share in the optimal basket

In a country which has the exchange rate pegged to the USD, EUR/USD volatility translates one-to-one into EUR/local currency changes. Lithuania used to be such a country, but stopped being one as soon as the reference currency for the currency board was switched to EUR. In the case of managed and floating exchange rates, exposure of the local currency to cross-exchange rate movements is much less obvious, making the tool useful.

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

109

Among the NMS, Romania and Latvia still exhibit significant attachment to the USD (see Figure 6.5). Thus the Dollar’s weakness against the Euro in 2003-2004 was translated into larger LTV/EUR and ROL/EUR depreciation. The sources of a large share of USD in the optimal basket could include dependence on raw materials trade, large USD-denominated debt, historical and cultural reasons, and the exchange rate regime. The first two causes are fundamentally justified and could call for some caution in establishing a very rigid exchange rate regime with respect to the Euro. The other two can quite safely be ignored, being endogenous and dependent either on the present or past currency regime13. Lithuania and Hungary managed to reduce significant shares of USD in their currency trading baskets to (close to) zero in less than a month. Lithuania achieved this by switching the currency board reference currency from USD to the Euro, while in Hungary it was enough to change the composition of the currency basket from 30% USD and 70% EUR (quite well reflected in the optimal basket calculations), to 100% Euro. This experience suggests that the introduction of any form of ERM-II could be sufficient to change trading habits. 6. CONCLUSIONS Assuming similar CPI inflation rates at the beginning of the ERM-II period, a wide band should make it easiest to meet the price stability criterion, followed by a narrow-band arrangement and pegs. The narrow-band ERM-II seems to be the most problematic in terms of exchange rate stability, followed by the wide-band ERM-II and currency boards. This classification, however, is not as clear-cut as the price-stability one. Much depends on the ERM-II parity level. Fixed exchange rate arrangements should be better for the budget deficit criterion, followed by the narrow and wide bands. Public debt statistics, however, could be aided by the nominal appreciation possibility, provided foreign debt is significant and exchange rate-inflation pass-through is not too fast. The current standing of the NMS suggests that rigid exchange rate regimes, including pegs and currency boards allow by far the fastest nominal convergence. The straightjacket they impose on fiscal policy, and the resulting smaller accumulation of public debt and currency stability are undisputed. Bond yields are not systematically higher in such countries, as the lack of nominal appreciation is compensated for by a lower short-term exchange rate risk. Finally, inflation, while theoretically suffering from the lack of a nominal appreciation buffer that compensates for relative price adjustments, does not seem to be as big a problem so far: on average inflation in currency board countries has been very close to that of the floating exchange rate countries. A fixed exchange rate does not solve the problem of final parity speculations fully. However, the countries with Euro-linked currency boards are especially likely to avoid a bond yield pick-up and the higher fiscal costs related to the risk of the final parity being set at a weaker level than that prevailing during the pre-ERM-II

110

MATEUSZ SZCZUREK

peg period. What is more, the unlikely switch from fixed exchange rates into a target zone with a weaker parity may still lead to nominal appreciation. The temptation to devalue for countries with other exchange rate regimes will be mitigated by political pressures (the required agreement of other MS to the final parity), by the fiscal costs of sterilized intervention defending the stronger edge of the band, by the potential inflationary impact of such a move, and by the the shortterm effect of foreign-currency denominated debt inflation, which could offset the reduction of the real value of local currency debt.

NOTES 1

Yet recent research suggests that the scale of the HBS effect is not significant, possibly thanks to high non-tradables’ productivity growth (see for example Blaszkiewicz, Kowalski, Rawdanowicz & Wozniak, 2005). 2 Delgado & Dumas (1992) show that for trended fundamentals (like trend money demand increase related to relative price adjustments) widening the currency band widens the bounds on fundamentals (the intervention limits) by three orders of magnitude. For mean reverting fundamentals the result holds only if the band is centered on the mean reversion point. 3 Other models of this kind are shown in Bertola & Svensson (1993). Flood & Garber (1992) extend the model to allow for discrete interventions. 4 See for example Garber & Svensson (1994) for the derivation of the solution, which uses Ito’s lemma 5 For other models of this kind see for example Drazen (1999) and Eichengreen, Rose & Wyplosz (1996). 6 Formally this is not excluded by the rules but in practice, how could one argue for a weaker conversion rate than parity, if the market rate had never been more than 2.25% above it in the preceding 2 years? 7 Real exchange rate appreciation, however, increases wealth, imports and consumption. As such, the losses in exporters’ CIT could well be compensated by higher VAT intake. 8 Much depends on the parity set, but there is no reason to believe the parity would be weaker for the band. 9 To be fair, Romania had 164% YoY CPI inflation in that period. 10 The old argument about sterilized interventions not being effective in countries with a fully opened capital account is important here. Taking this argument seriously, the sterilization costs would be infinitely high, assuming local currency short-term interest rates are higher than their foreign equivalents (portfolio balance arguments would ensure the capital inflow is not infinite). 11 Another option is to include a GDP growth differential which increases real money demand. 12 I am grateful to Bartosz Pawlowski for pointing to the simple analytical solution to the problem. 13 Yet abrupt change in volatility patterns created for whatever reasons could cause some problems for the real economy, creating unexpected currency risk. For example, if a dollarized property market suddenly becomes Euroized, initially optimal investment decisions like USD financing become suboptimal. USD debt, at first naturally hedged by the fact that the value of the real estate is also quoted in Dollars (sudden Dollar strength increases also the local currency value of the property), becomes unhedged when house pricing switches to local currency or the Euro.

REFERENCES Bertola, G. & Svensson, L.E.O. (1993). Stochastic Devaluation Risk and the Empirical Fit of Target-Zone Models. Review of Economic Studies, 60(3), 689-712, July. Blackwell Publishing. Blaszkiewicz, M., Kowalski, P., Rawdanowicz, L. & WoĨniak, P. (2005). Harrod- Balassa- Samuelson Effect in Selected Countries of Central and Eastern Europe. CASE Reports, 57. CASE - Center for Social and Economic Research. Buiter, W. & Grafe, C. (2002). Anchor, float or abandon ship: Exchange rate regimes for the Accession countries. EIB Papers, 7(2), 51-71.

EXCHANGE RATE REGIMES AND NOMINAL CONVERGENCE

111

Corsetti, G., Pesenti P. & Roubini, N. (1999). Paper tigers? A model of the Asian crisis. European Economic Review, 43, 1211-1236. Csermely, A. & Vonnak, B. (2002). The role of the exchange rate in the transmission mechanism in Hungary. Paper presented at the research meeting on ‘Monetary Policy Transmission in the Euro Area and the Accession Countries’, National Bank of Hungary, October. Djajiü, S. (1989). Dynamics of the exchange rate in anticipation of pegging. Journal of International Money and Finance, 8, 559-571. Delgado, F. & Dumas, B. (1992). Target zones, broad and narrow [in:] Krugman, P. & Miller, M. (eds.) Exchange Rate Targets and Currency Bands, Cambridge: Cambridge University Press. Drazen, A. (1999). Political Contagion in Currency Crises. NBER Working Paper, 7202 Eichengreen, B., Rose, A.K. & Wyplosz, C. (1996). Contagious Currency Crises: First Tests. Scandinavian Journal of Economics, 98 (4), 463-484. Flood, R. & Garber, P. (1984). Collapsing exchange rate regimes: some linear examples. Journal of International Economics, 17, 1-13. Flood, R. & Garber, P. (1992). The linkage between speculative attack and target zone models of exchange rates: some extended results [in:] Krugman, P. & Miller, M. (eds.) Exchange Rate Targets and Currency Bands, Cambridge: Cambridge University Press. Froot, K.A. & Rogoff, K. (1991). The EMS, the EMU, and the transition to a common currency. NBER Working Paper, 3684. Garber, P. & Svensson, L.E.O (1994). The Operation and Collapse of Fixed Exchange Rate Regimes. NBER Working Paper, 4971. Habib, M. (2002). Financial contagion, interest rates and the role of the exchange rate as a shock absorber in Central and Eastern Europe, Discussion Paper, 7/2002, Bank of Finland Institute for Economies in Transition (BOFIT). Ichikawa, M., Miller, M. & Sutherland, A. (1990). Entering a preannounced currency band, Economics Letters, 34 (4), 363-368. Elsevier. Krugman, P. (1979). A model of balance of payments crises. Journal of Money, Credit, and Banking, 11, 311-325 Krugman, P. & Rotemberg, J. (1992). Speculative attacks on target zones [in:] Krugman, P. & Miller, M. (eds.) Exchange Rate Targets and Currency Bands, Cambridge University Press, Cambridge. Miller, M. & Sutherland, A. (1992). Britain’s return to gold and entry into the EMS: joining conditions and credibility [in:] Krugman, P. & Miller, M. (eds.) Exchange Rate Targets and Currency Bands, Cambridge University Press, Cambridge. Obstfeld, M. (1994). The Logic of Currency Crises, Cahiers Économiques et monétaires, 43, p. 189, Banque de France. Obstfeld, M. & Stockman, A.C. (1985). Exchange-Rate Dynamics [in:] Kenen, P.B. & Jones, R.W. (eds.) Handbook of International Economics, 2, Amsterdam: North-Holland. Pelkmans, J., Gros, D. & Nunez Ferrer, J. (2000). Long Run Economic Aspects of the European Union’s Eastern Enlargement. Working Paper, 109. Scientific Council for Government Policy, WRR. Rodrik, D. & Velasco, A. (1999). Short Term Capital Flows. NBER Working Paper, 7364. Smith, G. W. & Smith, T. (1990) Stochastic process switching and the return to gold, 1925. Economic Journal, 100, 164-175.

LUKASZ W. RAWDANOWICZ

CHAPTER 7

EMU ENLARGEMENT AND THE CHOICE OF EURO CONVERSION RATES

1. INTRODUCTION In May 2004, ten new members states (NMS) joined the European Union (EU). At the same time they became members of the European Monetary Union (EMU) with a derogation. This implies that they are obliged to participate in the EMU at some point in the future. The standard accession path to the EMU requires staying in the exchange rate mechanism (ERM-II) for at least two years and the fulfillment of the Maastricht convergence criteria (stability of prices, exchange rates, convergence of long-term interest rates, and public debt and government budget deficit limits)1. At the end of June 2004 three NMS had already joined the ERM-II (Estonia, Lithuania and Slovenia). Under the ERM-II the central parity against the Euro is set and the domestic currency exchange rate must remain within a ±15% band around this parity. Any devaluation of the central exchange rate implies a new start for the reference period. The central exchange rate may be adopted as the irrevocable rate at which the currency is fixed in the EMU, though this rate could be changed during final negotiations. In this context the issue of choosing an appropriate final conversion rate arises. In 1999, the countries that took part in the introduction of stage III of the EMU fixed their currencies to the Euro at their ERM parities. These were initially set before 1979 and then devalued on several occasions. In the case of the UK, which joined the ERMI in 1990, the central parity was based on the purchasing power parity criterion (MacDonald, 2000). What should be the strategy for the NMS? A general view is that the central parity should reflect an equilibrium exchange rate (EER). However, putting this wisdom into practice is not straightforward, as no universal concept of EER exists. Different theoretical frameworks have different policy implications. In order to have a meaningful discussion about the choice of the appropriate exchange rate, equilibrium conditions must be clearly defined. Various issues relating to the choice of nominal conversion rates for the NMS upon entry into the EMU are investigated against this background. This chapter starts with a very short overview of exchange rate regimes in the NMS (section 7.2).

113 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 113-129. © 2006 Springer. Printed in the Netherlands.

114

LUKASZ W. RAWDANOWICZ

Then, a general background to the discussion on fixing nominal exchange rates (NER) in the EMU is outlined in section 7.3. Given that several conceptions of theoretical models of EER exist, a short overview of them is provided in section 7.4. Section 7.5 deals with the problems of testing and the practical implementation of these models. As the usefulness of theoretical EER models and the precision of their estimates are subject to certain reservations, other economic circumstances that should be considered when choosing the conversion rate are discussed in section 7.6. Finally, all the arguments are summarized and conclusions drawn in section 7.7. 2. EXCHANGE RATE REGIMES IN THE NMS Unfortunately, a clear-cut classification of exchange rate regimes poses a number of problems. This applies to a free-floating regime in particular, which in its pure form is rather rare in the real world. As Calvo & Reinhart (2002) suggest, many economies exhibit ‘fear of floating’ and in practice resort to intervention policies in order to steer the nominal exchange rate (NER). If this is the case, the classification problem boils down to a comparison of de facto and de jure exchange rate regimes. On the other hand, fixed peg exchange rate regimes are more transparent due to their clear institutional frameworks, in the form of either a currency board arrangement, adoption of a foreign currency (e.g. dollarization), or a conventional peg (to a foreign currency or a basket of foreign currencies). The NMS have adopted a wide range of different exchange rate regimes that, in some cases, underwent a significant evolution (see Table 7.1). At the beginning of transition in the early 1990s, when high inflation was pervasive and stabilization policies had to be put in place in many of the NMS, these economies resorted to fixed pegs. Later on, some of them gradually moved towards more flexible arrangements along the lines of adopting a direct inflation targeting (DIT) framework. This was the case in the Czech Republic, Hungary, Poland and Slovakia2. Poland could be classified as a free floater, whereas the Czech Republic and Slovakia are classed as managed floaters, since they have been intervening to curb excessive NER volatility. Hungary, on the other hand, unilaterally shadows the Euro. Unlike the above countries, Slovenia maintained a managed float from the beginning of the 1990s until joining the ERM-II in June 2004. The second distinctive group of NMS went for fixed peg regimes that have been sustained until today. This was the case in Estonia, Lithuania and Latvia. The former two have currency board arrangements that are maintained under the ERM-II as a unilateral commitment. Latvia currently has a fixed peg to the SDR. Malta and Cyprus, which do not come from a transition background, adopted conventional fixed peg regimes. The Maltese Lira is fixed to a basket consisting of the Euro, Dollar, and Pound Sterling, and Cyprus shadows the Euro.

THE CHOICE OF EURO CONVERSION RATES

115

3. EMU ENLARGEMENT AND CONVERSION RATES – THE BACKGROUND Before discussing the conversion rates at the time of EMU accession, several general remarks are required. They are intended to put the topic into a proper perspective. Table 7.1. Exchange rate regimes in the NMS Country/ currency name Cyprus Cyprus Pound (CYP) Czech Rep. Czech Koruna (CZK) Estonia Estonian Kroon (EEK) Hungary Hungarian Forint (HUF) Latvia Latvian Lat (LVL) Lithuania Lithuanian Litas (LTL) Malta Maltese Lira (MTL)

Poland Polish Zloty (PLN) Slovakia Slovak Koruna (SKK) Slovenia Slovenian Tolar (SIT)

Exchange rate regime Features/History ERM II compatible (unilateral commitment) Since 1999 central parity: 0.5853 CYP/EUR. History: peg since 1960 with a changing composition of the basket; in June 1992 CYP pegged to ECU with a ±2.25% band; in 1999 re-pegged to EUR; in January 2001 the band was widened to ±15%. Free float DIT. History: 1993-97 peg to a basket (USD and DEM); in 1997 financial crisis, devaluation and floating of CZK. ERM II with the currency board as a unilateral commitment Since 27 June 2004 central parity: 15.6466 EEK/EUR. History: currency board to DEM/EUR since 1992. ERM II compatible (unilateral commitment) Central parity: 282.63 HUF/EUR; DIT. History: 1989-95 peg to a basket (changing composition and gradual widening of the band); since May 1995 pre-announced monthly devaluation; in May 2001-June 2003 peg at 276.1 HUF/EUR with ±15% band. Peg to the SDR Since February 1994 peg at 0.7997 LVL/SDR with a ±1% band. ERM II with the currency board as a unilateral commitment Since 27 June 27 2004 central parity: 3.45280 LTL/EUR. History: currency board introduced in 1994; re-pegged from USD to EUR in February 2002. Peg to a basket Currency basket (EUR, USD, GBP) with a trading margin of ±0.25%. History: peg since 1964 with changing currencies and weights in the basket; since 1989 the basket has included EUC/EUR, USD, GBP); in August 2002 new weights introduced (EUR70%, GBP-20%, USD-10%). Free float DIT. History: 1990-91 peg to USD; since May 1991 peg to a basket (USD, DEM, GBP, FRF, CHF); since end-1991 crawling peg with a monthly devaluation rate; gradually lowering of the devaluation rate and widening of the band; since 1999 basket consists of USD and EUR and the band is ±15%; in April 2000 full floating of PLN. Managed float History: 1993-October 1998 peg to a multi-currency basket (from July 1994: USD and DEM), the band very narrow until 1995 and then gradually widened to 7% in 1997; October 1998 abandoning of the basket peg but interventions introduced. ERM II Central parity: 239.640 SIT/EUR. History: prior to ERM II entry (27 June 2004) managed float with a constant devaluation.

Notes: As of November 2004. Source: Author’s compilation based on information from national central banks.

First, joining the EMU is the final stage of economic integration into the EU. This integration involves the creation of a single market with the free movement of

116

LUKASZ W. RAWDANOWICZ

goods, capital and labor, as well as the adoption of a common currency and the pursuit of a single monetary policy. The process is aimed at bringing about stronger and more stable economic growth, improved economic efficiency, greater competitiveness, and some price level convergence. Thus, although the issue of selecting conversion rates may seem important, there are many other aspects of economic integration into the EMU that will determine its success. Given the ultimate long-term goal of improved economic performance, the importance of the choice of a particular level of conversion rate should not be overestimated. Second, membership of the EMU ultimately requires the NER to be fixed to the Euro. Changes in the NER are sensitive issues for policy makers given their impact on international competitiveness. This aspect usually dominates the whole discussion on exchange rates. Consequently, sectors directly exposed to international competition are implicitly the main focus3. Although the NER is only one of several determinants of international competitiveness (together with prices, wages, the price of other production inputs, productivity, and profit margins, etc.), it is often believed to be a quick and effective remedy for problems with external imbalances. Consequently, the NER, rather than other factors that are in actual fact central, is the main focus in political debates. Given some price and wage rigidities, this is to some extent justified, though adjustments in prices in response to NER changes should not be ignored. In theory, NER changes should be offset by corresponding changes in prices in the long run, leaving real exchange rates (RER) unchanged. Ignoring these adjustments could result in inappropriate policy conclusions. Although fixing the NER in the EMU is irrevocable, the choice of a particular conversion rate will by no means have an everlasting effect on the economy. Apparently, a change in the NER and in the exchange rate regime has far wider effects for an economy than for tradables’ competitiveness and national prices alone. In an era of largely liberalized capital flows, it has a potentially important impact on foreign currency denominated assets and liabilities. As the NMS have some debt denominated in foreign currencies, this effect should not be ignored4. In addition, the consequences for domestic price stability and investors’ confidence should also be taken into account. Third, while setting a Euro conversion rate in each NMS, interactions with other currencies should be analyzed. This refers to the notion of global consistency in exchange rate models, as discussed in Isard & Faruqee (1998) and Alberola, Cervero, Lopez & Ubide (1999) among others. A two-country framework with one bilateral exchange rate works well in a theoretical world, but in the real world, with many currencies, this may not be a good approach. Given this fact and the possibility that many of the NMS will join the EMU at relatively short intervals, factors such as the equilibrium between the Euro and the NMS’ currencies should be taken into account. If one focuses primarily on trade competitiveness, the number of currencies that should be incorporated into such an analysis could be selected based on the shares of trade with a country’s main trading partners. For most of the NMS, the EU-15 and the Eurozone are the main trading partners (see Chapters 1 & 2 of this volume). According to IMF data, in 2003 exports to the Eurozone ranged from roughly 23% of total merchandise exports in Cyprus to 65% in Hungary5. Similar

THE CHOICE OF EURO CONVERSION RATES

117

figures are also recorded on the import side. Consequently, it would be justified to focus only on mutual exchange rates among the NMS and the Eurozone. It should be stressed that setting the ‘internal’ exchange rate right (as between EMU member countries) in the enlarged Euro area does not require the Euro to be in equilibrium vs. non-EMU currencies6. These two aspects can be treated separately, and this chapter will deal only with the former. 4. THEORETICAL CONCEPTS OF THE EQUILIBRIUM EXCHANGE RATE AND EURO CONVERSION RATES The economic literature dealing with EER theories and assessment of the misalignment of various currencies, as well as empirical testing of these theories, has proliferated in recent years – see for instance Williamson (1994a), Allen & Stein (1995), Montiel & Hinckle (1999), MacDonald & Stein (1999), MacDonald (2000), and Isard, Faruqee, Kincaid & Fetherston (2001). In general, three approaches can be distinguished: finding the fundamental equilibrium exchange rate (FEER), the behavioral equilibrium exchange rate (BEER), and the natural real exchange rate (NATREX). The notion of the FEER, popularized by Williamson (1985), is based on the concept of internal and external macroeconomic equilibrium. The former is defined in terms of output at the full-employment level as well as low and sustainable inflation, whereas the latter is defined in terms of the sustainable and desired net flows of capital between countries that are internally balanced (Clark & MacDonald, 1999). The FEER indicates the exchange rate that would prevail under ‘ideal economic conditions’. Thus, this approach should be viewed as normative. It boils down to the calibration of the exchange rate with a set of well-defined economic conditions (Clark & MacDonald, 1999). The FEER approach states that a given equilibrium position should be viewed as ‘static’ (MacDonald, 2000). If it involves a stock-flow inconsistency, it cannot represent a true steady-state equilibrium. Wren-Lewis (1992) noted that the FEER approach implicitly assumes convergence of the actual real effective exchange rate with its FEER value. Thus, a medium-run current account theory of exchange rate determination is embedded in this approach. It simply assumes that any divergence in an exchange rate will be eliminated, although the adjustment process is not explicitly demonstrated. The misalignment based on the FEER model should be interpreted as a misalignment resulting from the departure of macroeconomic variables from their fundamental-equilibrium levels (defined in terms of the internal and external balance). In a sense, the FEER model points to an ideal situation of equilibrium in all markets. The BEER seeks an explicit relationship between macroeconomic fundamentals and the exchange rate. Fundamentals are based on a selected theoretical framework and practically any exchange rate theory can be embedded in this approach. Estimations of a BEER are usually done for a single-equation model, where the real exchange rate is explained using eclectic determinants – i.e. based on various

118

LUKASZ W. RAWDANOWICZ

exchange rate theories such as the balance of payments theory, the Harrod – Balassa - Samuelson (HBS) effect, uncovered interest parity, etc.7. The estimated BEER model provides information about the current misalignment. The latter term means a misalignment stemming from transitory and random effects, i.e. factors not treated as ‘fundamental’ determinants of the exchange rate (MacDonald, 2000). The BEER method also makes it possible to calculate a ‘fundamental’ EER and therefore the total misalignment (as with the FEER). This concept has been dubbed PEER (permanent equilibrium exchange rate), and should be treated as a way of calibrating the BEER at equilibrium values (MacDonald, 2000). The PEER decomposition allows for a gauging of the extent to which the misalignment implied by the BEER is permanent or transitory. For instance, MacDonald (2000) pursued such a decomposition using Gonzalo & Granger’s (1995) methodology, which makes it possible to extract a permanent component from a vector of cointegrated variables in the Johansen cointegration system. This method does not have sound theoretical foundations, and like the Hodrick-Prescott filter – which is commonly used to extract a permanent trend – is mechanical in nature. Another concept of EER is NATREX. It shares some common features with both the FEER and BEER models. This approach, initiated by Stein (1990) and discussed thoroughly in Allen & Stein (1995), focuses on the estimation of a medium and long-run real EER. The term ‘natural’ is motivated by the fact that the model assumes the ‘natural’ convergence of the exchange rate with its long-run equilibrium, i.e. it is a self-equilibrating framework. The EER is defined as the rate that ensures the balance of payments equilibrium in the absence of cyclical factors, speculative capital movements and changes in international reserves (Gandolfo & Felettigh, 1998). As in the FEER model, the underlying notion of equilibrium refers to internal and external equilibrium. However, in contrast to the FEER, the NATREX dynamic theoretical framework is stock-flow consistent, and a shift from medium to long-run equilibrium is explicitly modeled. The NATREX approach should be viewed as a generic framework within which various models could be embedded. A particular form of the model can be constructed to reflect the specific features of a given economy (Gandolfo & Felettigh, 1998). The NATREX models are usually built around four blocks: a production function, an investment function, a social saving function, and the balance of payments equation. The most common method of NATREX estimation is exactly the same as for BEER models, where the RER is regressed on selected fundamentals. In the case of the NATREX approach however, these fundamentals, which are derived explicitly, form the reduced-form equations of the underlying model8. Finally, the purchasing power parity (PPP) hypothesis should be mentioned. The PPP is a building block of many exchange rate models and has been extensively tested. The PPP paradigm assumes that the NER of any two currencies should closely reflect the relative purchasing powers of the two monetary units as represented by national price levels. The PPP model should be treated as a condition of international arbitrage rather than as an exchange rate determination theory per se (see below).

THE CHOICE OF EURO CONVERSION RATES

119

Having briefly surveyed selected concepts of EER, we assess their usefulness for selecting the Euro conversion rates upon EMU accession. The ‘fundamental’ concepts of EER (FEER, NATREX and PEER) assume explicitly that the rate is consistent with the internal and external equilibrium in the entire economy. They refer to the long-term EER, and thus may not be the most appropriate concepts for assessing a current misalignment and for providing practical guidelines for shortterm policy choices (see section 7.6). When choosing a Euro conversion rate, it could be argued that it is not justified to base this rate on long-term equilibrium concepts. Such benchmarks implicitly assume that the economy is at a long-term equilibrium (externally and internally balanced), and it would not make sense to expect the NMS to meet this condition when joining the EMU. Instead, in the context of setting the nominal parity, the EER should be considered as a level of the exchange rate that is consistent with other macro variables for a given point in time. In this respect, the ‘behavioral’ concepts of the EER seem to be more suitable, as they could be expected to identify an ‘equilibrium’ based on the current set of fundamental determinants. However, it appears that behavioral approaches are not ideal either. The assessment of a current misalignment depends on how well a given theory or set of theories capture exchange rate dynamics. For instance, a BEER model may indicate a misspecification or omitted variable problem rather than a misalignment stemming from random effects (i.e. everything that is not explained by the ‘true’ model)9. In practice, it is difficult to choose any particular form of the exchange rate determination model. However, if the estimated model demonstrates good explanatory (and possibly decent forecasting) properties, then the assessment should be less controversial. Finally, the conversion rates could also be based on the PPP framework. Although PPP was criticized as a good metric of exchange rate misalignment (Williamson, 1994b; MacDonald, 2000), it possesses some theoretical characteristics that make it attractive for this purpose. The PPP refers to the international arbitrage that equilibrates prices of tradables internationally. In a sense, it is a behavioral concept which reflects the current state of markets and does not refer to a theory-underpinned steady-state equilibrium. As was argued above, this feature makes it a good theoretical concept for selecting conversion exchange rates. However, it is for this particular reason that PPP has been disregarded as a ‘fundamental’ EER benchmark. 5. EQUILIBRIUM EXCHANGE RATE MODELS IN PRACTICE Policy recommendations concerning the appropriate level of the exchange rate require not only a proper and instructive theoretical framework, but also one that is operational. Therefore, this section discusses problems of the testing and the practical application of theoretical concepts of EER. In many papers in which real EERs are estimated, the real (effective) exchange rate based on consumer prices is used10. In such cases, the RER (q) could be disaggregated into the following components:

120

LUKASZ W. RAWDANOWICZ q = qT + ȕ(p*NT - p*T) - Į (pNT - pT)

(7.1)

The first term is the RER deflated using the prices of tradables only (pT and p*T, ‘*’ stands for a foreign country), and the latter two terms are the ratio of relative prices between the two countries (pT and p*T – are the domestic and foreign prices of tradables, and pNT and p*NT – of non-tradables, respectively). This disaggregation is derived from domestic and foreign inflation equations given by: p = (1-Į) pT + Į pNT and p* = (1- ȕ) p*T + ȕ p*NT

(7.2)

Į and ȕ denote the shares of non-tradable prices in the overall price index at home and abroad, respectively11. The disaggregation in equation 7.1 highlights two distinctive effects: the first concerns competitiveness in the tradable sector (as measured by qT), and the second focuses on the relative internal price ratio. If one is mostly interested in international competitiveness (as discussed in Section 7.3), then the analysis should mainly deal with qT. In the short run under a floating exchange rate regime, changes in the NER could be very abrupt and hard to explain. If price rigidities exist, large shifts in the NER could significantly impact on the economy. For instance, a significant nominal appreciation of the domestic currency may lead in the short run to, ceteris paribus, a contraction of profit margins, employment, wages or output. High volatility of the NER and price rigidities gave rise to the stylized fact that it is the key driver of changes in the RER under floating regimes. This is also the case for the NMS12. Thus, under free floating regimes it seems that the main problem with testing exchange rate models and explaining changes in RER lies in determining the NER, especially in the short and medium term. Against this background Meese & Rogoff’s (1983) empirical findings can be quite discouraging. They demonstrate that various structural exchange rate models do not have better out-of-sample forecasting properties than a simple random walk. The development of new models and estimation techniques in the 1980s and 1990s did not greatly change this finding. The structural models provided better forecasts only in the long term – longer than one year (Rogoff, 2001). On the other hand, under fixed exchange rate regimes the main challenge in understanding moves in RER lies in explaining movements in relative prices (particularly the relative prices of non-tradables and tradables). Given the above considerations, it can be argued that for countries with more flexible exchange rate regimes, an approximation of RER obtained by using relative prices is a very poor proxy13. In countries with a fixed exchange rate, there is a higher correlation between changes in relative prices and RER deflated by consumer prices (see Table 7.2). At this point it should be noted that at the beginning of the transition process in the NMS high inflation (both of tradables and non-tradables prices) was prevalent. The inflation was attributable to price liberalization and fiscal imbalances monetized by central banks. During this period, high inflation was the key driver of RER (both as defined in terms of the CPI, as well as in terms of prices of tradables only). Thus,

THE CHOICE OF EURO CONVERSION RATES

121

the real appreciation of exchange rates at the beginning of transition could be a separate phenomenon from what we are examining at present14. Table 7.2. Correlation coefficients between the Euro RER and relative prices Flexible exchange rate regimes Country (period) CZE (94:1-03:4) 0.008 HUN (96:1-03:4) 0.062 POL (94:1-03:4) 0.422 SLK (96:1-03:4) 0.461 SLO (94:1-03:4) -0.083

Fixed peg regimes Country (period) EST (94:1-03:2) 0.898 LAT (94:1-03:2) 0.641 LIT (94:1-03:2) 0.682

Notes: Correlations are based on annual growth rates for quarterly data. The Euro RER is defined as the Euro NER multiplied by Eurozone inflation and divided by domestic inflation. Relative prices are defined as a ratio of domestic to Eurozone relative prices (non-tradables vs. tradables). Tradable prices are defined in terms of producer prices in manufacturing and non-tradable prices as consumer prices of services. Cyprus and Malta are omitted due to missing data. Source: Author’s calculations base on IFS-IMF, OECD and national sources.

In light of these general conceptual problems with estimating EER models, the operational problems of particular models are addressed below. Estimations of BEER are highly data demanding, with regards to both the coverage of the data and length of time series. The estimation of these models boils down to finding long-run elasticities. Given the common presence of non-stationary time series, these elasticities are estimated in the cointegration framework (either in VAR or single equation models). The models are usually estimated using annual or quarterly time series. A comparable and consistent data set for the NMS is available only from the beginning of the 1990s, and if one wants to include several explanatory variables, the econometric robustness of such estimations can be questioned. In addition, unit root tests suffer from the problem of weak power when the tested variables are close to having a unit root and when structural breaks are present in the series (Maddala & Kim, 1998). This leads to a further problem if one has to deal with a mixture of stationary and non-stationary variables in the cointegration analysis. The problem of short time series for the estimation of EER models could be circumvented by the application of panel techniques. Panel estimations have recently become a workhorse of empirical tests in international economics. As indicated in a survey by MacDonald (1998), panel estimations tend to result in better results in terms of their economic and statistical properties than single-equation estimations do. Recently, several methods of estimation of dynamic panel models have been popularized, in particular panel Dynamic OLS (Mark & Sul, 2002; Kao & Chiang, 2000), Pooled Group Mean Estimator (Pesaran, Shin & Smith, 1999), panel mean group Fully Modified OLS (Pedroni, 2001). These methods were employed for the estimation of EER for selected NMS in Kim & Korhonen (2002) – PMGE, MacDonald & Wojcik (2003) – DOLS, and Rahn (2003) – FMOLS. Although panel models give more econometrically robust results, they tend to downplay country-specific factors. Consequently, they are better suited to providing

122

LUKASZ W. RAWDANOWICZ

evidence for some general economic model, rather than providing precise guidelines on a given variable (like an exchange rate level) for a particular country. In addition, the problems of stationarity and cointegration are further complicated in comparison to the time-series models (Pesaran, 2000; Maddala & Kim, 1998; Pedroni, 1999). Unit root tests cannot always be interpreted meaningfully. The usual null hypothesis of the panel unit root tests is the joint non-stationarity of RER in the panel. Therefore, a rejection of the null hypothesis could mean that only one of the tested series is stationary (Pesaran, 2000). The estimation of a simple PPP model also poses practical problems. The empirical literature on PPP is vast15. The general consensus on PPP says that it is a long-run phenomenon. Usually, time series of price indices are used in PPP estimations, as price level data are not easily available. Different models (the absolute vs. the relative PPP hypothesis) and tests (either testing the stationarity of RER or estimating coefficients for the relationship between the NER and the corresponding price indices for the home and the foreign country) have been pursued. The most commonly quoted obstacles to testing the PPP hypothesis and explanations for the failure to prove it empirically, stem from the distinction between tradables and non-tradables (which is far from clear in practice), the existence of tariff and non-tariff trade barriers, monopolistic practices for pricing to segmented markets, imperfectly competitive markets where changes in prices are costly, transport costs, and differences in indirect taxes (e.g. Cecchetti, Mark & Sonora, 2000). In addition, PPP testing could be affected by the distinction between exchange rate regimes. Given that PPP refers primarily to international arbitrage and not to the determination of NER per se, it seems more reasonable to find evidence for PPP under fixed exchange rate regimes than under free floats. The arbitrage is less likely to occur when changes in the NER are very volatile and unpredictable, since an international comparison of prices is more difficult and the arbitrage is more risky16. In this respect, a given market’s power to affect international prices should also be mentioned. Given the small shares of the NMS in European markets, one should not expect the impact of domestic prices in the NMS to have a large effect on prices in the Eurozone. But the opposite causality seems to be more likely. This point also relates to the issue of exchange rate pass-through. In general, theoretical and empirical evidence (relating to a more pervasive local currency pricing mechanism in economies characterized by markets with monopolistic competition and a higher share of non-tradables in the structure of the economy and consumption) suggest a weaker pass-through (from changes in NER to inflation) in developed economies as compared to that in developing countries. Disregarding the problems of time-series tests of PPP, a comparison of price levels could be useful in order to get some benchmarks for the final conversion rates for NMS currencies into Euro. Data on comparable price levels are collected by Eurostat. However, in practice it is difficult to decide what basket of goods and services should be used for comparison. This choice is crucial for the calculation of PPP exchange rates (see Table 7.3). The PPP exchange rates differ greatly depending on the particular category of goods and services being used. Moreover, data on price levels are only available with a significant time lag and are very often

THE CHOICE OF EURO CONVERSION RATES

123

subject to revision. This makes the PPP method unattractive for timely policy recommendations. It is interesting to note that for almost all countries and categories of goods and services, the PPP exchange rates were more appreciated than market exchange rates. In general, goods that seem to be more tradable (for instance, durable goods or machinery and equipment) are closer to market exchange rates as opposed to the PPP exchange rates for services. In addition, there is a positive relationship between the percentage distance of PPP exchange rates for goods from the market exchange rate and the analogous distance for PPP exchange rates for services. This pattern could reflect the contribution of non-tradable prices to the prices of tradables, as the production and sales/delivery of tradables are very likely to contain a services component17. In countries that have lower relative costs of services, one could expect lower (observed) prices of tradables – if observed tradable prices are believed to include the cost of distribution services etc.18. Such a mechanism is proven empirically by Lee & Tang (2003) for 12 OECD countries. Based on Table 7.3, it seems plausible that the mechanism is also present in the NMS. In general, it should be noted that the PPP exchange rates are likely to be downward biased (i.e. likely to indicate overly appreciated exchange rates) as a benchmark for conversion rates, although little can be said precisely about the magnitude of this effect19. Table 7.3. PPP exchange rates, 1999 (national currency per Euro) Item Market exchange rate in 1999 Total goods Consumer goods Non-durable goods Semi-durable goods Durable goods Capital goods Food and non-alcoholic beverages Alcoholic beverages, tobacco and narcotics Clothing and footwear Furnishing, household equipment, routine household maintenance Transport Machinery and equipment Total services Consumer services Government services Collective services Individual services

CZE 36.83 21.71 22.20 20.96 22.64 26.84 21.69 18.30

HUN 252.65 156.01 154.27 147.33 148.35 192.44 162.80 137.06

POL 4.23 2.66 2.69 2.52 2.71 3.56 2.63 2.34

SLK 44.07 23.77 22.87 21.82 22.44 28.91 25.66 22.50

EST 15.64 10.21 9.58 9.12 10.65 10.42 11.81 9.68

LAT 0.62 0.43 0.42 0.38 0.55 0.51 0.45 0.42

LIT 4.26 2.69 2.63 2.46 2.97 3.08 2.96 2.56

SLO 193.65 157.41 163.45 165.98 157.15 159.70 151.85 186.66

MAL 0.425 0.359 0.375 0.371 0.340 0.425 0.338 0.384

CYP 0.578 0.460 0.510 0.476 0.483 0.682 0.385 0.489

EMU 0.939 1.062 1.084 1.132 1.119 0.932 1.030 0.990

23.94 156.48 2.96 21.61 10.83 0.54 3.16 139.57 0.602 0.641 0.820 21.65 144.96 2.50 20.15 10.26 0.54 2.82 144.81 0.330 0.456 1.296 20.71 144.63 2.56 21.87 9.16

0.38 2.37 131.56 0.371 0.466 0.949

21.87 30.12 9.82 11.03 8.54 10.78 7.31

0.45 0.54 0.17 0.21 0.12 0.14 0.11

189.36 205.47 73.09 87.05 60.35 74.27 51.29

2.82 3.58 1.40 1.62 1.16 1.32 1.04

20.23 35.50 8.68 9.95 7.55 9.06 6.75

9.35 14.32 4.58 5.48 3.66 3.98 3.39

2.48 3.58 1.00 1.17 0.83 0.97 0.74

155.36 181.70 103.51 112.89 91.83 96.83 88.15

0.429 0.434 0.29 0.37 0.21 0.22 0.20

0.492 0.564 0.38 0.39 0.38 0.34 0.41

1.076 0.966 0.771 0.784 0.753 0.803 0.698

Notes: Exchange rates in domestic currency per 1 EUR but in case of EMU it is the EUR/USD rate. Source: PPP exchange rates – Eurostat and OECD, market exchange rates (national currency per Euro, annual average) – IFS, IMF.

Finally, mentioned, conversion model or

the practical drawbacks of ‘fundamental’ models of EER should be although these models are argued to be unsuitable for setting the Euro rates from a philosophical point of view. A calibration of the FEER an estimation of NATREX and PEER models require arbitrary

124

LUKASZ W. RAWDANOWICZ

assumptions, and are subject to uncertainty given problems with the definition of variables and finding corresponding real data. In particular, the FEER models require the use of (non-observable) potential output and, as noted by Bayoumi, Clark, Symansky & Taylor (1994) and the IMF (1998), plausible estimates of FEER may vary quite substantially20. As far as NATREX models are concerned, the most common way of testing them exhibits the same problems as the estimation of BEER models. 6. GUIDELINES FOR CHOOSING EURO CONVERSION RATES The above discussion demonstrates that despite the existence of several different theoretical EER approaches, not all of them are well suited for making inferences about the conversion rates upon EMU accession. Additionally, all of these concepts suffer from important operational problems that cast doubt on the precision of their point estimates of EER. This critique does not mean, however, that all these models and their estimations should be discarded entirely. Some useful information can still be elicited, if interpreted carefully21. What are the other considerations that should be taken into account and that can help in selecting the conversion rates? As pointed out by the IMF (1998), a complete assessment of the exchange rate misalignment should not be based on model estimates only, but should also take into account a broader range of macroeconomic aspects like policy-mix, structural factors, stage of development, and so on. For instance, if one is interested primarily in international competitiveness then a more detailed analysis of this issue could be conducted. This would involve an investigation of the different determinants of wages, productivity, prices of other production inputs, as well as other potential factors driving NER – capital inflows, interest rate differentials, etc. Disregarding problems with the empirical measurement of EER and ensuing misalignments, another practical issue arises. What should be done if a misalignment is diagnosed? How could equilibrium be achieved and what are the implications for the choice of Euro conversion rates? Not all of the theoretical models give clear answers to these questions. In the case of BEER models, the current misalignment is defined in terms of model residuals, which may stem either from ‘temporary’ shocks or misspecification of the model. For ‘fundamental’ NATERX models, it seems that nothing can be done as the long-run equilibrium is achieved ‘naturally’ (i.e. via defined interactions with other variables in the model), unless long-term fundamental determinants are changed. In the case of FEER, no dynamics are explained and no clear implications on the adjustment to equilibrium can be found. It should also be mentioned that as fundamentals and some exogenous parameters in the BEER and NATREX models change, the long-run equilibrium could change as well. Thus, reaching exchange rate equilibrium may be like chasing a constantly moving goal. Given these considerations, setting the nominal Euro parity based on the equilibrium value does not solve the problem of reaching this equilibrium. Shifting the NER may not be enough to achieve internal and external equilibrium for the

THE CHOICE OF EURO CONVERSION RATES

125

entire economy22. Although one can expect the NER to have some impact on fundamentals like productivity, saving and investment ratios, consumption, and public debt, pinning down the exact channels in a dynamic environment with price adjustments and other shocks to the economy is very difficult. Besides, one can expect some lags in the convergence to equilibrium. Only in the PPP model can the adjustment of the NER at a given point in time, restore equilibrium unambiguously. While deciding on a change in the NER upon accession to the ERM-II/ EMU, the consequences of this change should be addressed. In this context, a potential nominal shock to the economy and reactions of financial markets need to be discussed. The first issue refers primarily to the effects on trade flows, firms’ profits, employment, etc. According to conventional macroeconomic analysis, too strong an exchange rate spurs recessionary effects and, conversely, an undervalued currency produces expansionary ones. It is often the case that not only is the mere fact of misalignment important, but also its magnitude. One could expect a non-linear relationship between the NER and real performance23. However, price adjustments and the revaluation of foreign currency denominated assets and liabilities should also be taken into account. Understanding this problem is very important in the context of EMU membership. The NMS will have to meet the Maastricht criteria, so when setting conversion rates, macro objectives other than simply correcting the NER misalignment must be considered (for instance inflation, interest rates, and debt targets). It could be argued that in the NMS, the effects of choosing an ‘excessively strong’ and an ‘excessively weak’ conversion rate are asymmetric. Given the importance of export markets for the economy of each NMS on the one hand and for the Eurozone on the other, as well as the potential power to affect international prices, the consequences of weak NMS conversion rates seem to be far less acute for the Eurozone than for the NMS24. Regarding the potentially higher inflation induced by the depreciation of a domestic currency (an ‘excessively weak’ conversion rate), this should not pose a major threat if countries do not suffer from significant macroeconomic imbalances and pursue stable and low-inflation oriented policies (this is currently the case in many NMS). Nonetheless, more simulations of such effects for each country would be needed. An investigation of these issues could also be augmented by an assessment of the current phase of the business cycle and expectations of monetary conditions in the EMU during the ERM-II period and after EMU accession. For instance, there would be less of a rationale to opt for a weak conversion rate if a country is in a boom and Eurozone interest rates are expected to be less restrictive than domestic rates from the perspective of this country. Financial market reactions and expectations should also be considered. The minimum two-year period in the ERM-II leaves some scope for an adjustment of the NER. If the desired NER (in terms of the central parity and ultimately of the irrevocable conversion rate) differs from the current level of the exchange rate, then a credible announcement of the central parity in the ERM-II may allow for a gradual convergence to this rate. As Reluga & Szczurek (2002) point out, it is very likely that the market exchange rate will converge with the announced central parity if this announcement is credible. The credibility is strengthened by the announcement of

126

LUKASZ W. RAWDANOWICZ

the final EMU conversion rate, which should take place in advance of entry into the EMU25. The nominal convergence of exchange rates was clearly visible for example in Spain and Portugal prior to their accession to the EMU. However, if financial markets realize that the choice of a weak exchange rate is simply a way to deal with other structural problems in the economy, then the resulting undermining of the confidence of foreign investors could do more harm than good26. An example is the Hungarian case of the devaluation of the Forint’s central parity in June 2003. This move was intended to relieve domestic manufacturers from appreciation pressure. However, the loss in competitiveness was not only attributable to the strong NER of the Forint, but also to buoyant growth in wages. In addition, at that time a loose fiscal policy was pursued, which fuelled the growth of wages, household consumption and in turn inflation and the current account deficit. After the realignment of the Forint’s central parity against the Euro on June 4, 2003 (by 2.3% of the initial parity), the Forint depreciated considerably. When the Forint approached 270 HUF/EUR, the National Bank of Hungary decided to raise interest rates by 100 basis points to 7.5%. This measure did not have the expected effect and monetary policy was tightened again on June 19, 2003 (by 200 basis points). In the aftermath, market yields and the Forint exchange rate started to stabilize. Summarizing, the realignment of the central parity took place at the expense of higher interest rates, and did not solve the problem of deteriorating international competitiveness as wages continued to increase at a fast pace. 7. CONCLUSIONS This chapter deals with the choice of nominal Euro conversion rates in the NMS upon their accession to the EMU. It is stressed that the level of a nominal peg is only one of many aspects of EMU enlargement and its consequences should be relatively short-lived. Issues of trade competitiveness are usually the main concern behind NER adjustments, although other determinants of international competitiveness should not be ignored. A general view is that the central parity of the ERMII should reflect an equilibrium position. It is argued that despite the existence of several EER theories, not all of them are useful for indicating the nominal conversion rates of NMS currencies to the Euro. The selection of a particular theoretical model determines the interpretation of any potential exchange rate misalignment and its policy implications. It is claimed that the BEER and PPP models seem to be more appropriate theoretical constructs for the selection of conversion rates, but their practical implementation is complicated by many operational problems (as is the case with other models). This makes them less attractive for the purpose of guiding practical policy decisions. The problems of choosing an appropriate theoretical EER framework and the intrinsic uncertainty about particular estimates lends support to the view that the range of potentially appropriate conversion rates is quite wide and other issues must be taken into account. Concerns about international competitiveness should be augmented with a broader macroeconomic assessment, especially in view of the need to meet the Maastricht criteria.

THE CHOICE OF EURO CONVERSION RATES

127

For the financial markets it is not only the level of the conversion rate that matters, but also how and in what circumstances the rate is selected. A smooth transition to the Euro and the minimization of potential shocks to the economy should be the key concern behind the choice of the conversion rate. In this respect, recommendations for the selection of nominal conversion rates should be dependent on the current exchange rate regime. Given the lack of precise theoretical and empirical guidelines coming from EER models, the countries with fixed exchange rate regimes (Estonia, Lithuania, Latvia and Malta) seem to be better off leaving their current fixed rates unchanged. It seems that it is precisely these considerations that were indeed applied for the choice of the ERM-II “central parity” in Estonia and Lithuania. This recommendation is not as straightforward, however, for Latvia and Malta. These countries peg to a basket of currencies. Consequently, at some point this basket should be changed to contain only the Euro. Similarly, in countries that currently have ERM-II compatible exchange rate regimes (Hungary and Cyprus), the central parity should not be changed, unless there is sound evidence that the nominal exchange rate is the factor impacting negatively on macroeconomic performance. The choice for countries with more flexible exchange rates regimes (Poland, the Czech Republic and Slovakia) is more problematic as there is no existing nominal reference point. For these countries a more detailed and country-specific analysis of the issues discussed in this paper should be undertaken. In particular, the consequences of changes in the NER on trade competitiveness, inflation, and debt repayments should be investigated. Similar matters were considered when setting the ERM-II central parity in Slovenia.

NOTES 1

The exact language of the Maastricht criteria and a recent assessment of NMS can be found in ECB (2004). 2 This section describes exchange rate regimes as of the end of November 2004. 3 These are the tradables, i.e. manufactured goods, some services that are easily traded internationally, and also agricultural commodities. Interestingly, depending on the definition, in 2003 tradables in NMS comprised on average around 30% of total GDP. 4 According to joint BIS-IMF-OECD-World Bank statistics, in 2002 the total foreign debt to GDP ratio (at current prices converted at market exchange rates) ranged from around 19% in the Czech Republic to 158% in Malta. The corresponding figures for the foreign debt with a maturity of one year or less range between 6.7% in Slovenia and 86.6% in Malta. 5 Data is derived from the IMF database – Direction of Trade Statistics. It refers to trade in goods only. Thus, trade links for Cyprus and Malta, which mainly provide tourism, are biased downwards based on these statistics. 6 Conversion rates could be treated as the ‘internal’ EER for a block of countries. The ‘external’ EER of the Euro vs. other currencies must incorporate information on EMU trading partners. 7 The lack of a clearly defined underlining theoretical model and an ad hoc selection of explanatory variables in some applications of BEER models has been criticised by Stein (1999). Examples of the empirical application of this approach include Baffes, Elbadawi & O’Connell (1997) or Clark & MacDonald (1999).

128

LUKASZ W. RAWDANOWICZ

8

There have also been attempts to estimate all equations of the model simultaneously (Gandolfo & Felettigh, 1998). 9 Estimates of eclectic BEER models pose less risk of an omitted variable problem than estimations of one specific and narrowly defined model of exchange rate determination (such as uncovered interest rate parity or PPP). A similar point is made by Stein (1999). 10 For instance, MacDonald & Wojcik (2003) and Alberola, Cervero, Lopez, & Ubide (1999). 11 For the sake of simplicity, the shares are implicitly assumed to be equal in many applications. 12 For a set of charts demonstrating this phenomenon see Rawdanowicz (2003). 13 The relative prices are most commonly explained by the HBS effect. However, demand variables or endowments in factors of production play a role in some models as well. 14 The appreciation of the real exchange rates in the CEE economies at the beginning of transition is investigated in Grafe & Wyplosz (1997). 15 See for example Rogoff (1996), Pedroni (2001), Moon & Perron (2002), Bayoumi & MacDonald (1998). 16 For instance, Parsley & Wei (1996) find a positive relation between deviations from PPP and NER volatility. 17 In the context of exchange rate models this phenomenon is discussed by MacDonald & Ricci (2001) and Lee & Tang (2003). 18 The mathematical proof of this hypothesis can be found in Rawdanowicz (2003). Price levels collected in PPP surveys in the joint Eurostat-OECD project refer to final consumer prices, and thus should be expected to contain a services component (for instance, retail trade mark-ups, transport costs, etc.). 19 For the tests of the relative PPP model and the hypothesis of a non-tradable component in NMS tradeable prices see Rawdanowicz (2004). 20 See Rawdanowicz (2002) for operational problems with the estimation of FEER. 21 For a survey of empirical estimations of EER for selected NMS see Egert (2003). 22 For instance, some theoretical models – like the EER rate model – imply the authorities cannot influence RER by changing a NER. 23 For further discussion of these issues see Collins & Razin (1997). They demonstrate empirically that only a very high overvaluation leads to slower GDP growth, and medium and high undervaluation to higher growth for a large sample of developed and developing countries. 24 According to the IMF data, in 2003 merchandise exports of the NMS to the Eurozone accounted for between 2% (in Cyprus) and 44% (in Slovakia) of GDP, whereas the exports of the Eurozone to the NMS constituted less than 2% of the GDP of the former. The figure for Cyprus is downward biased as it does not include trade in services. 25 EMU entry takes place usually at the beginning of a year. Thus, depending on the start of ERM-II membership, there could be a couple of months between actual entrance into the EMU and the end of the two-year reference period when the final decision is made. 26 The mechanism here could be compared to the second generation model of a currency crisis.

REFERENCES Alberola, E., Cervero, S.G., Lopez, H. & Ubide, A. (1999). Global equilibrium exchange rates: Euro, Dollar, "ins", "outs" and other major currencies in a panel cointegration framework. IMF Working Paper, WP/99/175, December. Allen, R.P. & Stein J.L. (1995). Fundamental Determinants of Exchange Rates. Oxford: Clarendon Press. Baffes, J., Elbadawi, I.A. & O’Connell, S.A. (1997). Single-Equation Estimation of the Equilibrium Real Exchange Rate. World Bank Country Economics Department Paper, 1800. Bayoumi, T., Clark, P., Symansky, S & Taylor, M. (1994). The Robustness of Equilibrium Exchange Rate Calculations to Alternative Assumptions and Methodologies. [in:] Williamson, J. (ed.). Estimating Equilibrium Exchange Rates, 19-59. Washington: Institute for International Economics. Bayoumi, T. & MacDonald, R. (1998). Deviations of Exchange Rates from PPP: A Story Featuring Two Monetary Unions. IMF Working Paper, WP/98/69, May. Calvo, G.A & Reinhart, C.M. (2002). Fear of floating, The Quarterly Journal of Economics, 117(2), 379408, May.

THE CHOICE OF EURO CONVERSION RATES

129

Cecchetti, S., Mark, N. & Sonora, R.J. (2000). Price level convergence among United State cities: Lessons for the European Central Bank. NBER Working Paper, 7681, May. Clark, P.B. & MacDonald, R. (1999). Exchange Rates and Economic Fundamentals: A Methodological Comparison of BEERs and FEERs, [in:] MacDonald, R. & Stein, J. L. (eds.). Equilibrium exchange rates, Kluwer Academic Publisher. Collins, S.M. & Razin, O. (1997). Real exchange rate misalignments and growth. NBER Discussion Paper, 6174, September. ECB (2004). Convergence Report 2004. Frankfurt: European Central Bank. Egert, B. (2003). Assessing Equilibrium Exchange Rates in CEE Acceding Countries: Can We Have DEER with BEER without FEER? A Critical Survey of the Literature. Focus on Transition, 2, 38106, Oesterreichische Nationalbank. Gandolfo, G. & Felettigh, A. (1998). The NATREX: an Alternative Approach - Theory and Empirical Verification, C.I.D.E.I. Working Papers, 52, http://www.eco.uniroma1.it/cidei/wp/cidei52.pdf Gonzalo, J. & Granger, C.W. (1995). Estimation of Common Long-Memory Components in Cointegrated Systems. Journal of Business Economics and Statistics. 13 (1), 27-35. Grafe, C. & Wyplosz, C. (1997). The Real Exchange Rate in Transition Economies. CEPR Discussion Papers, 1773, December. IMF (1998). World Economic Outlook. Ch. III – The Business Cycle, International Linkages, and Exchange Rates. International Monetary Fund, May, 55-73. Isard, P. & Faruqee, H., eds. (1998). Exchange Rate Assessment. Extensions of the Macroeconomic Balance Approach. IMF Occasional Paper, 167, December. Isard, P., Faruqee, H., Kincaid, G.R. & Fetherston, M. (2001). Methodology for Current Account and Exchange Rate Assessments. IMF Occasional Paper, 209, December. Kao, C. & Chiang, M.H. (2000). On the Estimation and Inference of a Cointegrated regression in Panel Data. [in:] Baltagi, B.H. (ed.) Nonstationary Panels, Panel Cointegration, and Dynamic Panels. Advances in Econometrics, 15, Elsevier Science. Kim, B.Y. & Korhonen, I. (2002). Equilibrium exchange rates in transition countries: Evidence from dynamic heterogeneous panel models. BOFIT Discussion Papers, 15/2002. Lee, J. & Tang, M.K. (2003). Does Productivity Growth Lead to Appreciation of the Real Exchange Rate? IMF Working Paper, WP/03/154, July. MacDonald, R. (1998). What Do We Really Know About Real Exchange Rates? Oesterreichische Nationalbank Working Paper, 28. MacDonald, R. (2000). Concepts to Calculate Equilibrium Exchange Rates: An Overview. Economic Research Group of the Deutsche Bundesbank Discussion Paper, 3/00, July. MacDonald, R. & Ricci, L. (2001). PPP and the Balassa Samuelson Effect: The Role of the Distribution Sector. IMF Working Paper, WP/01/38, March. MacDonald, R. & Stein¸ J.L., eds. (1999). Equilibrium Exchange Rates. Kluwer Academic Publisher.

JOANNA SIWINSKA AND PIOTR BUJAK

CHAPTER 8

THE SHORT-RUN MACROECONOMIC EFFECTS OF DISCRETIONARY FISCAL POLICY CHANGES

1. INTRODUCTION The adoption of a common currency in EU countries has revived interest in the influence of fiscal policies on aggregate demand and output. However, despite the efforts of many researchers, the short- as well as long-run effects of fiscal policy remain unclear. The standard Keynesian or ‘conventional’ view of fiscal policy holds that in the short run, fiscal adjustments (expansions) reduce (stimulate) aggregate demand and due to sticky wages, prices or other market rigidities, these demand shifts affect production and output factors. This proposition has been challenged by the permanent income life cycle hypothesis (PILCH) and its generalization, the RicardoBarro Equivalence Theorem, which holds that the level of aggregate demand is unaffected by the tax/debt mix or by permanent changes in government spending; the former leaves household consumption decisions unaffected, while the latter crowds out. Fiscal policy is thus ineffective in stimulating or dampening the output, at least in the short run1. ‘Real life’ has, however, demonstrated that there are other options as well, and that the Keynesian and Ricardian theories have not exhausted all the possible effects of fiscal policy. It has been well documented that fiscal policy can have ‘nonKeynesian’ effects, i.e. fiscal expansion can cause a recession, and fiscal retrenchments may trigger expansion in economic activity. These effects do not fit into either the Keynesian or the Ricardian traditions, and although a considerable amount of research has been done concerning these phenomena, they remain largely unexplained. The main goal of this chapter is to reassess the short-run effects of discretionary fiscal policy, concentrating on the response of consumption to fiscal policy changes. We use two data sets: a sample of data from OECD countries that has already been used in previous studies and additional data from transition economies. The chapter is structured in the following way: section 8.2 briefly outlines the main theories regarding the short-run influence of fiscal policies on economic activity, concentrating on possible explanations of the non-Keynesian effects and

131 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 131-145. © 2006 Springer. Printed in the Netherlands.

132

JOANNA SIWINSKA AND PIOTR BUJAK

reviews the empirical literature. Section 8.3 is empirical – we run several regressions for both OECD and transition economies to test the proposition of non-linear effects of fiscal policy further. Section 8.4 contains our conclusions. The results of our research support the proposition of possible non-linearity in consumer behavior. We find that in times when the fundamental fiscal position of a country is sound, fiscal policy has short-run Keynesian effects: fiscal expansion stimulates private consumption while fiscal contraction dampens it. In times of fiscal distress, for example when there are doubts over the solvency of the government, these effects are reversed: loose fiscal policy dampens private consumption and vice versa. Thus households behave in a nonlinear fashion – in ‘normal fiscal times’ they are ‘Keynesian’, while in ‘bad fiscal times’ they exhibit non-Keynesian behavior. A similar non-linear effect is also present in the transition economies. 2. THE SHORT-RUN EFFECTS OF FISCAL POLICY ON OUTPUT This section briefly outlines the main competing theories regarding the influence of fiscal policies on consumption and aggregate demand2. The simplest Keynesian view assumes that fiscal policy influences aggregate demand in a straightforward fashion: fiscal expansion stimulates consumption and demand, while fiscal contraction reduces them. In the short run, output follows aggregate demand changes due to a positively sloped aggregate supply curve. This is what Elmendorf & Mankiw (1998) call the ‘conventional view’ of fiscal policy. Approaches based on rational, forward-looking consumers and PILCH challenge these conventional predictions. A ‘logical completion’ of PILCH is Ricardian equivalence (Seater, 1993, p. 143). According to this approach, when government spending is fixed, the tax/debt mix is irrelevant for consumption decisions; it does not change the path of private consumption, the only effect being the altered composition of national savings. This does not imply that fiscal policy as a whole does not influence consumption, since what matters is the level of the government’s use of resources, i.e. government spending. Provided government spending does not enter into consumers’ utility function, a permanent rise in public expenditure means an equal decrease in permanent income, hence any increase in public spending is offset by a decrease in private consumption (Seater, 1993; Barro, 1989). A temporary increase in public spending, holding its permanent level fixed, will not change the consumption pattern, which may imply a temporary increase in aggregate demand3. The assumption of a forward-looking consumer, who is not restricted by liquidity constraints, and who maximizes his and his descendents’ lifetime utility renders the Keynesian theory of fiscal policy almost completely irrelevant. Of course, if we relax any of these assumptions the results of the PILCH/Ricardo view change – they become more Keynesian, in the sense that fiscal policy has Keynesian effects on consumption. It is obvious that the actual impact of fiscal policy on aggregate demand and output cannot be settled solely on theoretical grounds, as there are too many unresolved questions concerning the crucial assumptions mentioned above. It can

THE EFFECTS OF DISCRETIONARY FISCAL POLICY CHANGES

133

only be established empirically. The seminal paper by Campbell and Mankiw (1990) provides empirical evidence for consumption not following a random walk, as implied by PILCH, but being ‘excessively sensitive’ to predictable changes in income and ‘excessively smooth’ to surprise changes in income; thus ‘Keynes’s original consumption function starts to look more attractive’ (Mankiw, 2002, p. 456).

There has also been a considerable amount of research devoted to testing Ricardian equivalence. Although the conclusions are not uniform (Seater, 1993; Elmendorf & Mankiw, 1998), it seems that ‘most economists today agree with David Ricardo and doubt that Ricardian equivalence describes actual consumer behavior4’ (Elmendorf & Mankiw, 1998. p. 43).

Although Ricardian equivalence/PILCH might give an approximation of consumer behavior, these concepts do not describe economic reality fully and hence many fiscal policy actions do in fact influence aggregate demand and short-run economic activity in a Keynesian fashion. This state of the art was challenged in the early 1990’s, as neither of these theories – Keynesian or PILCH/Ricardian – could explain the phenomenon first described in a seminal paper by Giavazzi & Pagano (1990, p.81), who analyzed the effects of fiscal policy in Denmark and Ireland in 1982 and 1987-89 respectively and concluded that these were ‘…the two most striking cases of expansionary (fiscal) stabilizations in Europe’.

Since then, extensive research has aimed to explain this paradox. Some of the studies point to the possible non-linear reaction of consumption to fiscal policy changes as an explanation for this phenomenon. Among them are the models by Blanchard (1990), Bertola & Drazen (1993), Perotti (1999), and Sutherland (1997). All four models are neoclassical (i.e. based on PILCH) and each holds that expectations of future fiscal policy changes can give rise to non-linear consumer behavior5. Blanchard’s (1990) model is based on the notion that increases in tax rates lead to distortions in the economy. He assumes a critical level of taxation t*, such that distortions caused by taxes exceeding this level imply a decrease in output. Consequently, there is also an associated critical level of debt d* that implies a future tax rate above the critical level t* and thus a lower output. If consumers anticipate that this critical level of debt, d*, will be reached, then a fiscal consolidation that stabilizes or lowers the value of debt will allow the economy to escape from the expected highly distortionary tax trap. Therefore, expected permanent income and current consumption rise. In other words, today’s tax increase, which does not exceed the critical value t*, allows a larger future tax increase, and hence an output decrease, to be avoided. As a result, fiscal consolidation in ‘bad fiscal times’ can be good news and can raise consumption. Blanchard notes that if consumers have a constant probability of death (i.e they are not ‘Ricardian’), then in ‘normal times’ (i.e. when the economy is far from the critical debt level), fiscal policy will have Keynesian effects despite the neoclassical

134

JOANNA SIWINSKA AND PIOTR BUJAK

structure of the model. The model thus shows that consumers behave in a non-linear fashion – in ‘normal times’ they behave in a ‘Keynesian’ way (provided they have a finite horizon), in ‘bad times’ their behavior is reversed, which gives rise to nonKeynesian effects. The model of Bertola and Drazen (1993) analyzes the effects of public expenditure on consumption. It assumes that government spending follows a random walk with a positive drift. To satisfy the budget constraint, the government is forced, at some point in time, to lower expenditure. The model assumes that these fiscal stabilizations are triggered by given levels of government spending, a lower value, g*, or a higher value, g**. The lower value, g*, triggers a stabilization (a decrease in spending to a new level g) with probability p* which is less than one. If g* is surpassed, stabilization occurs with probability equal to one at the higher value, g**. In ‘normal fiscal times’, i.e. far from g*, an increase in public expenditure usually lowers private consumption because infinitely-lived consumers correctly recognize that today’s increase in public spending means increased future taxation and lower permanent income. However, the offset in consumption expenditure is not equal to the increase in the deficit because agents know that sometime in the future the increase in government spending must be reversed. More interesting is the behavior of consumption around the ‘trigger values’ of government spending. When government spending approaches g*, a discrete cut to g is possible but not certain (note that this possibility has been included in consumers’ estimate of permanent income). If the trigger value is reached and public spending is not cut, then consumption falls discretely as consumers revise their expectations of permanent income. Thus the non-linearity of the model emerges – consumption falls significantly and discretely when government spending increases by a small amount, provided that a fiscal retrenchment was expected but has not materialized. Another non-linearity is present around the second ‘trigger value’, i.e. when the level of public expenditure is close to g**. Agents know that stabilization will inevitably take place soon and increase their consumption, anticipating a lower value of future taxes. When the stabilization does occur, consumption is already higher (at a level corresponding to the smaller public spending), and thus a large cut in government spending leaves consumption unchanged. Yet another model has been provided by Sutherland (1997). In this model, the government implements a restrictive fiscal policy in the form of a large tax hike when public debt reaches critical levels. At low values of debt the probability of fiscal contraction is very low. Therefore, a fiscal transfer from the government to households produces a Keynesian effect, since the probability that the future necessary increase in taxes will fall on the current consumers is low. The higher the debt value is, i.e. the closer it is to the trigger point of fiscal stabilization, the greater is the fall in consumption resulting from further increases in fiscal deficit. This is because consumers know that there is a high probability that the stabilization – a large tax increase –- will take place during their lives and that the future income loss will be much higher than today’s government transfer. Again, the non-linearity in consumer behavior gives rise to non-Keynesian effects. It is worth noting that the model’s results depend crucially on the assumption of a finitely-lived consumer. The

THE EFFECTS OF DISCRETIONARY FISCAL POLICY CHANGES

135

non-linearity diminishes when consumers have infinite lives, in which case the results are Ricardian. A further influential model has been developed by Perotti (1999). Like Campbell and Mankiw (1990), Perotti assumes that consumers are not homogenous: one group is rational and forward-looking, while the other group is either liquidity-constrained or myopic, so the consumption function of the latter group is purely Keynesian. Perotti also implicitly assumes that an increase in government spending stimulates pre-tax income, while taxes dampen it. The distortions caused by taxes increase in a non-linear fashion; because policymakers do not smooth taxes out, future taxes are expected to be higher than at present. The non-Keynesian effects of fiscal policy in the model emerge when expectations about the future disposable income of rational, unconstrained consumers change as a result of an unexpected change in public spending or taxes, mainly through the alteration of the future tax path6. Note that ‘Keynesian’ consumers have a constant propensity to consume out of current income, and the bigger their proportion in society, the less likely is it for the non-Keynesian effects to appear. An unexpected increase in public expenditure may exert a positive or negative effect on the consumption of the ‘PILCH’ consumers, depending on the strength of the output stimulus relative to the negative effect of the implied (future) tax increase. If the output stimulus outweighs tax distortions, consumption will increase as a result of the government spending shock, while if the opposite is true, consumption will decrease. Therefore, an increase in government spending may have perverse demand effects, provided the share of the ‘PILCH’ consumers is not too low and that the effect of the (expected) tax distortions outweighs the positive effects of spending. The latter effect is more likely when the present discounted value (PDV) of the government’s financing needs is large – for instance, when the government is struggling with mounting debt. Similarly, a tax increase may have different effects depending on the expected path of future taxation. An unexpected tax increase today may stimulate consumption among forward-looking, unconstrained individuals, as it implies significantly lower tax distortions in the future. Again, the emergence of a nonKeynesian reaction by consumers depends on the share of unconstrained consumers and on the PDV of the financing needs of the government. In ‘bad fiscal times’ the PDV of the government’s financing needs is large, so today’s tax increase implies a significant decrease in future distortions. Perotti (1999) therefore argues that when the PDV of future government financing needs is low, fiscal policy will exert the usual Keynesian effects due to the reaction of liquidity-constrained consumers (provided the fiscal action is not expected). When, however, the PDV of financing needs is high, it is possible that the unexpected fiscal policy will have perverse effects, because the ‘PILCH’ consumers expect a future adverse change in income. The empirical literature, starting with the seminal contribution of Giavazzi and Pagano (1990) provides evidence for non-Keynesian effects, but the conclusions on what factors trigger these effects are mixed (see Table 8.1).

136

JOANNA SIWINSKA AND PIOTR BUJAK

Perotti (1999) finds evidence in favor of non-linear effects of fiscal policy and shows that high or growing public debt triggers them. A similar, although weaker, conclusion has been reached by Bhattacharya (1999). Pozzi (2001) provides additional evidence, in the spirit of Perotti’s work, based on data from Canada and Italy. Giavazzi, Japelli & Pagano (2000) fail to find excessive debt as significant in triggering non-linear effects. In their estimation the relevant factor is the size of the adjustment/ expansion: they find that large and persistent fiscal policy changes trigger the non-linearity. However, Cour, Dubois, Mahfous & Pisani-Ferry (1996) fail to find any strong relationship between the size of fiscal expansion/contraction and non-Keynesian effects. Alesina & Ardagna (1998) provide evidence that the composition of fiscal policy is the factor that causes the non-Keynesian effects. Similarly, Alesina & Perotti (1997) argue that the composition of fiscal adjustment is crucial for its macroeconomic effects. Pozzi, Heylen & Dossche (2002) test the influence of rising government debt on the proportion of liquidity-constrained consumers and find that as the debt increases so do liquidity constraints7. Although they do not explicitly test for the non-linear effects of fiscal policy, their evidence suggests that the proportion of ‘Keynesian’ consumers increases in ‘bad fiscal times’, which should make the emergence of nonKeynesian effects less likely in those times8. Summing up, the effects of fiscal policies are unclear both from the theoretical and empirical work. There are three distinctive theoretical views on the short-run consequences of discretionary government actions: first, the Keynesian, or conventional, view that says that fiscal expansions or adjustments have symmetrical effects on demand and output. Second, the PILCH/Ricardian view, holds that fiscal policy is largely irrelevant, and the third view – the non-Keynesian approach – proposes that fiscal policy can have perverse effects due to non-linearities. The impact of fiscal policy may be Keynesian or Ricardian (depending on assumptions) in good times, but in ‘bad fiscal times’, or in the case of certain kinds of fiscal policy composition, its effects will be reversed compared to conventional expectations. 3. EMPIRICAL EVIDENCE Our empirical research focuses on consumption behavior9. Using data on OECD as well as transition economies, we verify whether there is non-linearity in consumption induced by fiscal events. Note that non-linearity does not necessarily imply a non-Keynesian output effect: as it may be too weak to offset the other effects of fiscal policy, it simply signals that the overall impact of fiscal policy might not be uniform. Our estimation is based on the methodology of Perotti (1999) and Campbell and Mankiw (1990), that has been followed, among others, by Pozzi, Heylen & Dossche (2002). Following Perotti (1999), we model consumption changes as a function of expected disposable income changes and unexpected fiscal policy changes. The

THE EFFECTS OF DISCRETIONARY FISCAL POLICY CHANGES

137

expected income changes will lead to a change in the consumption of liquidityconstrained or myopic consumers. Unexpected fiscal policy shocks will cause both groups of consumers to adjust their consumption path. The expected changes will not affect the consumption of unconstrained, forward-looking consumers, and the reaction of myopic consumers to expected policy changes is already embedded in the coefficient of disposable income changes. Since we are interested in the reaction to discretionary policy changes, we use cyclically adjusted fiscal impulses. The cyclical adjustment also allows us to avoid problems caused by endogeneity of the fiscal variables. The fiscal impulse is measured as the change in the difference between personal income together with social security taxes and primary government expenditures. We estimate the following consumption function:

'cit = 'yit +'balit +'balit*D1it + eit

(8.1)

where: 'cit – consumption change; 'yit – expected change in disposable personal income; 'bal – unexpected, cyclically adjusted fiscal impulse, defined as personal income taxes and social security contributions minus primary expenditures; D1 dummy variable that equals 1 when the government debt in a country exceeds its mean plus one standard deviation computed for the period 1970- 2001 (i.e. the mean and standard deviation are computed separately for each country). To check the robustness of our results, we also employ a different definition of this variable denoting it as D1WS. This dummy equals one when the government debt in a country exceeds the mean plus one standard deviation computed for the whole sample. We also include a dummy variable D2 to incorporate the possible effect of the Stability and Growth Pact (SGP); D2 equals 1 for countries that have adopted the SGP. Fiscal rules should make the non-Keynesian effects more pronounced – a fiscal rule that requires a reversion of policy when a deficit and/or debt exceeds a certain threshold might make an otherwise myopic consumer more aware of government budget constraints. Secondly, the adoption of fiscal rules makes a fiscal contraction more probable in the short time horizon (the issue of time horizon is discussed in Sutherland’s model). We test the influence of the SGP by multiplying the budget balance variable by two dummies: D1 (or D1ws), which, as before, equals one in ‘bad fiscal times’ and D2 or (1-D2). All the variables are in real per capita terms; consumption and income are expressed as log differences; the change in budget balance is, following Perotti (1999), expressed as the real per capita change in the budget balance divided by the real per capita disposable income (this is done to take into account the significant cross-country differences in the size of the budget deficit). All data are from the OECD database. Before we actually estimate equation 8.1, we must compute the expected changes in disposable income and the unexpected changes in fiscal policy. Following Perotti (1999), these are calculated using the following system of VAR equations:

138

JOANNA SIWINSKA AND PIOTR BUJAK q

Xt = D +

¦ EX t i + Xt

(8.2)

i 1

where: Xt = ('ct, 'yt, 'balt). The estimated error terms represent the unexpected variations in variables, while the estimated dependent variables are the expected changes. We estimate equation 8.1 on panel data from OECD countries (all, except Poland, the Czech Republic and Hungary) for the years 1970-2001 using a simple within-effects estimator, which is a consistent estimation procedure in the case of generated regressors (see Perotti, 1999; Pozzi, 2001). Table 8.2 summarizes our results. The coefficient signs in estimation 1, without any dummies, are consistent with the Keynesian view. However, the inclusion of the dummy variable for ‘bad fiscal times’, i.e. large government debt (estimations 2 and 3) uncovers the non-Keynesian effect: in ‘bad fiscal times’ an increase in government spending dampens consumption, as in Perotti (1999). Hence, in times of fiscal distress households change their behavior. In estimations 4 and 5 we test for the effect of the SGP. The variable 'balit*D1*D2 (which means that a country has adopted the SGP and was experiencing ‘bad fiscal times’) is statistically insignificant, the variable 'balit*D1ws*D2 (which has similar meaning, but D1ws is computed differently than D1) performs better. Its coefficient is larger in absolute value than the coefficient of the variable 'bal*D1*(1-D2) (this variable means that a country is experiencing bad fiscal times but has not (yet) adopted the SGP), which might suggest that the adoption of the SGP makes the perverse effect of fiscal policy stronger. This is an interesting result, which should be tested in a few years’ time when we have more data points available10. We also test for the presence of a non-linear effect of fiscal policy in transition economies. Data availability allows us to include only Belarus, Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Kazakhstan, Latvia, Lithuania, Moldova, Poland, Romania, Russia and Slovakia in our sample. The earliest data point is the year 1991 and the latest is 2001. Due to the short time dimension of the data, we omit the VAR estimation, thus neither the changes in income nor those in fiscal shock are divided into expected and unexpected components. All variables are in real per capita terms. The budget balance variable is scaled by real per capita GDP, while consumption and income are given in log differences. The fiscal data are from IMF Government Finance Statistics, while the consumption and GDP variables are calculated from World Bank Development Indicators. Since we do not have data for disposable income, we proxy it by GDP. The budget balance data are for consolidated central government and are not cyclically adjusted. Due to the unavailability of a full dataset on government debt, we use budget deficits to proxy for ‘bad fiscal times’; thus times of fiscal distress for a given country i (dummy D1it=1) are defined as the mean plus one standard deviation of the budget deficit for country i over the relevant time period11.

THE EFFECTS OF DISCRETIONARY FISCAL POLICY CHANGES

139

Estimation of a consumption function that employs GDP and unadjusted fiscal data poses many econometric problems, one of them being the endogeneity of the right hand side variables. To (partly) account for this problem, we employ the twostep instrumental variable estimation technique12. The results shown in Table 8.3 suggest that in the transition economies, a nonlinear reaction by consumers to changes in fiscal policy might also be present. Note however, that the coefficient of 'bal*D1 (that is the influence of budget balance in times of fiscal distress) is smaller in absolute value than the coefficient of 'bal, which means that fiscal expansion in ‘bad fiscal times’ increases consumption but this effect is much weaker than in ‘good fiscal times’. Thus, it seems that consumers in transition economies do behave in a non-linear fashion, but the non-Keynesian effect does not outweigh the Keynesian response, so on average there are no perverse reactions of consumption to fiscal policy. This result is not surprising – as Perotti (1999) argues, the non-Keynesian effect depends, among other things, on the proportion of myopic and liquidity-constrained consumers. In transition economies liquidity constraints are probably more pronounced due to less developed financial markets. Moreover, the proportion of myopic consumers may be also greater. The citizens of formerly centrally planned economies may still not be familiar with the concept of government budget constraints and may not understand that the resources of the government are limited. These results must, of course, be treated with skepticism, due to short time period concerned and the poor quality of the data. 4. CONCLUSIONS The short-run effects of fiscal policy, despite the efforts of many researchers, are still a matter of controversy. The compromise position held by most economists is that the strict PILCH/Ricardian view, although theoretically interesting, is empirically fairly irrelevant; so fiscal policy will have Keynesian effects in the short run. These simple predictions have been shaken by observed episodes of the perverse effects of fiscal policy – so called ‘non-Keynesian’ effects. These effects might be caused, among other reasons, by non-linear consumer behavior that changes in times of fiscal distress. The main goal of our analysis was to test for the existence of this non-linearity. We found evidence that consumption reacts in a non-linear fashion to discretionary fiscal policy changes. The results of our estimations show that in ‘normal fiscal times’ consumers behave in Keynesian fashion. In ‘bad fiscal times’ their behavior tends to change – an expansionary fiscal policy lowers consumption, which is definitely not a Keynesian result. These results are not novel; they confirm the empirical conclusions reached by Perotti (1999) and are in line with the theoretical models of Sutherland (1997) and Blanchard (1990). We also found that these non-Keynesian effects may be reinforced by the SGP (although the very limited number of data points available do not allow us to reach definite conclusions). This result is especially appealing, as it

140

JOANNA SIWINSKA AND PIOTR BUJAK

reinforces the validity of the proposition on which the models of Blanchard (1990), Bertola and Drazen (1993) and Sutherland (1997) are based, namely that expectations of future fiscal policy changes influence today’s consumer behavior13. We have also tried to verify the presence of non-Keynesian effects in transition economies. As usual in these countries, a short time period and poor quality of the data are the reasons why estimation results must be treated with considerable caution. Nevertheless, our results suggest that, similarly to OECD countries, in transition economies the consumption function does not react in a linear fashion to changes in fiscal policy. When the fiscal position is sound, fiscal expansion stimulates consumption; this ‘conventional’ effect is also present in ‘bad fiscal times’ but is much weaker – the increase in consumption in such cases is close to zero. This result is consistent with Perotti’s (1999) model. Perotti argues that such a perverse effect of fiscal policy is dependent on the proportion of myopic and liquidity-constrained consumers – the bigger the proportion, the less likely the nonKeynesian effect. In transition economies this share is probably larger than in OECD economies for two reasons. First, financial markets are less developed, so it is probable that more consumers are liquidity-constrained, and, second, more consumers might be myopic as a consequence of many years of central planning and a lack of familiarity with the operation of the government budget constraint in a market-economy setting. The proportion of ‘Keynesian consumers’ will probably have a tendency to fall over time, and therefore the transition economies are likely to see the non-Keynesian effects of fiscal policy growing in future. Furthermore, our findings are not very good news for those EU politicians who would like to stimulate the economy through fiscal expansion. Expansionary fiscal policy, especially in the countries that have adopted the Euro, may not always have the usual effects.

NOTES 1 In the long run, the composition and magnitude of government spending and taxes, as well as the amount of public debt, may exert a significant impact on the growth rate. 2 A comprehensive review can be found in Hemming, Kell & Mahfouz (2002). 3 Consumption plans may be affected by public spending not only through the budget constraint but also through their utility, provided that government consumption enters into private utility functions. Depending on whether public consumption is a substitute or complement for private consumption, its increase will either decrease or increase private consumption. 4 Ricardo himself did not belief in Ricardian equivalence (Elmendorf & Mankiw, 1998). 5 These four models emphasize the demand-side sources of the non-Keynesian effects of fiscal policy. Some studies (see for example Alesina, Ardagna, Perotti & Schiantarelli, 1999) argue that the main reason for non-Keynesian effects lies on the supply side. 6 The ‘PILCH’ consumers do not react to expected changes in fiscal policy. Consumption only changes when fiscal policy is unexpected. 7 They argue that as public debt increases, banks may reduce the amount they lend to households. 8 Note that a non-Keynesian effect is more probable when consumers are not liquidity-constrained. 9 We follow the conclusion of Cour, Dubois, Mahfous & Pisani-Ferry (1996) that consumption seems to be a good candidate to explain non-Keynesian effects. Of course, the existence of non-Keynesian effects on the demand side does not exclude the possibility of supply side effects and vice versa.

THE EFFECTS OF DISCRETIONARY FISCAL POLICY CHANGES

141

10

Assuming that the SGP will survive. We ran a regression for the very limited sample of countries for which we had public debt data (i.e. D1=1 when the public debt of a country is excessively large), but then the deficit variable turned out to be not statistically significant. 12 We also ran a regression with dummy variables for well-known output shocks (the Russian crisis, Bulgarian crisis, the Balkan war, etc.). The dummy was not statistically significant. 13 As far as we know, the influence of the SGP has not so far been described in the literature on nonKeynesian effects. 11

REFERENCES Afonso, A. (2001). Non-Keynesian Effects of Fiscal Policy in EU-15. Working Papers, 2001/07. Department of Economics, Institute for Economics and Business Administration (ISEG), Technical University of Lisbon. Alesina, A., Ardagna, S., Perotti, R.. & Schiantarelli F. (1999). Fiscal Policy, Profits and Investment. NBER Working Paper, 7207. Alesina, A. & Ardagna, S. (1998). Tales of fiscal adjustment. Economic Policy, 27. Alesina, A. & Perotti, R. (1997). Fiscal Adjustment in OECD Countries: Composition and Macroeconomic Effects. IMF Staff Papers, 44 (2), 210-248. Barro, R. (1989). The Neoclassical Approach to Fiscal Policy. [in:] Barro (ed). Modern Business Cycle Theory. Harvard University Press. Bertola, G. & Drazen, A. (1993). Trigger Points and Budget Cuts: Explaining the Effects of Fiscal Austerity. American Economic Review, 83 (1), 11-26. Bhattacharya, R. (1999). Private Sector Behavior and Non-Keynesian Effects of Fiscal Policy. IMF Working Paper, WP/99/112. Blanchard, O. (1990). Comment on Giavazzi and Pagano. [in:] Blanchard, O. & Fischer, S. (eds.). NBER Macroeconomics Annual, 5, 111-122. Cambridge, Mass: The MIT Press. Blanchard, O. & Perotti, R. (1999). An empirical characterization of the dynamic effects of changes in government spending and taxes on output. NBER Working Paper, 7269. Cambell, J.Y. & Mankiw, N.G. (1990). Permanent Income, Current Income and Consumption. Journal of Business & Economic Statistics, 8 (3), 265-79, July. Cour, P., Dubois, E., Mahfous, S. & Pisani-Ferry, J. (1996). The cost of fiscal retrenchment revisited: How strong is the evidence? Serie des Documents de Travail, G9612, Institut National de la Statistique et des Etudes Economiques, Paris. Elmendorf, D. & Mankiw, G. (1998). Government Debt. NBER Working Paper, 6470 Fatas, A. & Mihov, I. (2001). The effects of fiscal policy on consumption and employment: Theory and Evidence. mimeo, August, http://faculty.insead.fr/mihov/files/FPandConsumptionAug2001.pdf Giavazzi, F., Japelli, T. & Pagano, M. (2000). Searching for Non-linear Effects of Fiscal Policy: Evidence from Industrial and Developing Countries. NBER Working Paper, 7460, January. Giavazzi, F. & Pagano, M. (1995). Non-Keynesian effects of fiscal policy changes; international evidence and the Swedish experience. NBER Working Paper, 5332, November. Giavazzi, F. & Pagano, M. (1990). Can Severe Fiscal Contractions Be Expansionary? Tales of Two Small European Countries. [in:] Blanchard, O. & Fischer, S. (eds.). NBER Macroeconomics Annual, 5, 75116. Cambridge, Mass: The MIT Press. Hemming, R., Kell, M. & Mahfouz, S. (2002). The Effectiveness of Fiscal Policy in Stimulating Economic Activity – A Review of the Literature. IMF Working Paper, WP/02/208. Hemming, R., Mahfouz, S. & Schimmelpfennig, A. (2002). Fiscal Policy and Economic Activity During Recessions in Advanced Economies. IMF Working Paper, WP/02/87. Lane, P. & Perotti, R. (1999). The Importance of Composition of Fiscal Policy: Evidence from Different Exchange Rate Regimes. Trinity College Dublin CEG Working Papers, 200111, Trinity College, Dublin, Economics Department. Mankiw, N.G. (2002). Macroeconomics (5th ed.). Worth Publishers McDermott, J. & Wescott, R. (1996). An Empirical Analysis of Fiscal Adjustment. IMF Working Paper, WP/96/59

142

JOANNA SIWINSKA AND PIOTR BUJAK

Miller, S.M. & Russek, F.S. (2003). The Relationship between Large Fiscal Adjustments and Short-Term Output Growth under Alternative Fiscal Policy Regimes. Contemporary Economic Policy, 21(1), 4158, January. Perotti, R. (2002). Estimating the effects of fiscal policy in OECD countries. Economics Working Papers, 015, European Network of Economic Policy Research Institutes. Perotti, R. (1999). Fiscal policy in good times and bad. The Quarterly Journal of Economics, 114(4), 1399-1436, November. Pozzi, L. (2001). Government debt, imperfect information and fiscal policy effects on private consumption. Evidence for two high debt countries. Working Paper, 2001/125. Faculteit Economie & Bedrijfskunde, Universiteit Gent. Pozzi, L., Heylen, F. & Dossche, M. (2002). Government debt and excess sensitivity of private consumption to current income: an empirical analysis for OECD countries. Working Papers of Faculty of Economics and Business Administration, 02/155, Ghent University, Belgium. Seater, J. (1993). Ricardian Equivalence. Journal of Economic Literature, 31 (1), 142-90, March. Sutherland, A. (1997). Fiscal Crises and Aggregate Demand: Can High Public Debt Reverse the Effects of Fiscal Policy? Journal of Public Economics, 65 (2), 147-162.

THE EFFECTS OF DISCRETIONARY FISCAL POLICY CHANGES

143

APPENDIX 8.1. TABLES Table 8.1. Review of the empirical literature on non-Keynesian effects Authors

Method of research

Sample

Results. Evidence of non-linear consumer behavior around large changes in primary structural budget balance; thus the size of adjustment may trigger non-Keynesian effects, but the magnitude of reversal is too weak to explain non-Keynesian effects. Evidence of non-Keynesian effects. Size of adjustment and composition are the factors that seem to influence the likelihood of their emergence. Non-Keynesian effects are possible. Increases in primary government spending reduce profits and investment: the strongest effect is caused by the government wage bill; increases in taxes reduce profits and investment, but the magnitude is smaller. Composition of fiscal policy is thus crucial for non-Keynesian effects. Evidence of Non-Keynesian effects. Composition of fiscal adjustment is an important factor in triggering these effects. Evidence of non-linear behavior – households move from non-Ricardian to Ricardian behavior as government debt reaches a threshold; debt level and debt history is important for the effects of fiscal policy. Keynesian results, but the multipliers are small; increases in T and G both have strong negative effects on investment spending (Keynesian theory predicts that increasing G should increase investment). Consumption is crowded in by increases in spending, but exports and imports are crowded out. Statistical analysis: evidence of nonKeynesian effects; large fiscal adjustments and expansions seem to be less Keynesian; and anti-Keynesian episodes are specific to large-scale fiscal episodes. Regression analysis: consumption behavior does account for non-Keynesian effects, however it is difficult to identify the factors that may be the cause of this non-linearity

Afonso (2001)

Regression analysis, consumption function

EU-15

Alesina & Ardagna (1998)

Statistical analysis; Probit estimation

OECD

Alesina, Ardagna, Perotti & Schiantarelli (1999)

Regression analysis; dependent variable: investment, profits

OECD

Alesina & Perotti (1997)

Statistical analysis

OECD

Bhattacharya (1999)

Regression analysis for each country and panel data approach; dependent variable - consumption

OECD

Blanchard & Perotti (1999)

VAR; effects of fiscal variables on GDP

US, postwar period; quarterly data

Cour, Dubois, Mahfous, & Pisani-Ferry (1996)

Statistical and regression analysis; dependent variable: consumption

OECD, 197195

Fatas & Mihov (2001)

VAR

USA

Giavazzi, Jappeli & Pagano (2000)

Regression analysis; dependent variable: national saving

18 industrial Evidence of non-linear effects of fiscal countries 1970policy; the factor explaining non-linearity is 96, 150 the size of fiscal adjustment/expansion industrial and

Keynesian results

144

JOANNA SIWINSKA AND PIOTR BUJAK developing countries 196095 (WB dataset)

Giavazzi & Pagano (1995)

Regression analysis; dependent variable: consumption

19 OECD countries, 1970-92

Giavazzi & Pagano (1990)

descriptive and regression analysis (consumption)

Denmark and Ireland

Hemming, Mahfouz & Schimmelpfennig (2002)

Authors analyze recession episodes only; statistical and econometric analysis; dependent variable: depth of recession

29 advanced economies, 1970-99; 61 recession episodes

Lane & Perotti (1999)

Regression analysis; dependent variable: employment, real value added and profitability in manufacturing (supply side)

Miller & Russek (2003)

Regression analysis: St. Luis equation, consumption equation and growth accounting equation

McDermott & Wescott (1996)

statistical and regression analysis of fiscal adjustments

Perotti (2002)

VAR

Perotti (1999)

Regression analysis, dependent variable:

OECD, 196494

Evidence of non-linearity: non-Keynesian reaction of consumption function emerges when changes in cyclically adjusted deficits are particularly large and persistent; size of fiscal adjustment matters. Non-Keynesian effects of fiscal consolidations; evidence of consumption boom when government cuts spending. Statistical analysis: fiscal expansions are more effective (i.e. have Keynesian effects), when: (i) debt is low; (ii) govt. is big; (iii) are expenditure based. Estimation: fiscal policies have Keynesian effects in closed economies, non-Keynesian effects in open economies. Government size matters – big governments have more Keynesian effects (because of larger automatic stabilizers). Negative effect of government wage spending on output, profitability and employment in the traded goods sector; composition of fiscal policy matters for its effect

St Luis equation: evidence of nonlinear effects around large and persistent fiscal expansions. Consumption equation: nonlinear behavior around large fiscal expansions, but not large enough to explain non-Keynesian effects, these effects are OECD magnified by the inclusion of ‘trigger points’ i.e. dummies for exceptionally high and rapid growth of government consumption and primary structural deficit. Growth accounting equation: no nonlinearity found. Non-Keynesian effects emerge during successful fiscal consolidations (i.e. those resulting in at least 3 percentage points reduction in public debt to GDP in the OECD 1970-95 second year after the adjustment has ended). The success of consolidations depends on its size (bigger ones being more successful) and on composition: spending cuts more successful Overall small multipliers; smaller (and in 5 OECD some countries negative) multipliers in the countries post-1980 sample, as compared to pre-1980. 19 OECD Evidence of non-linearity: non-Keynesian countries 1965- effects emerge when government debt is 94 high

THE EFFECTS OF DISCRETIONARY FISCAL POLICY CHANGES

Pozzi (2001)

Pozzi, Heylen & Dossche (2002)

consumption Regression analysis: consumption function Regression analysis: consumption function

145

semi-annual data for Italy and Canada 1960/1-1997

Evidence of non-linearity: when debt ratio increases, tax shocks have non-Keynesian effects

OECD, 1990-9

High government debt increases the share of credit-constrained consumers.

Table 8.2. Estimation results; OECD countries; dependent variable: 'cit Sample (adjusted): 1973-2001; Total panel (unbalanced) observations: 360 Estimation 1 Estimation 2 Estimation 3 Estimation 4 Estimation 5 0.67 0.66 0.65 0.67 0.65 'yit (9.30) (9.15) (9.11) (9.15) (9.12) -0.21 -0.34 -0.44 -0.34 -0.43 'balit (9.30) (-2.30) (-2.47) (2.30) (-2.43) 0.43 'balit*D1 (1.57) 0.44 'balit*D1ws (1.77) 0.40 'balit*D1 (1.40) *(1-D2) 0.61 'balit*D1 *D2 (0.93) 0.31 'balit*D1ws*(1(1.48) D2) 0.80 'balit*D1ws *D2 (1.58) R-squared: R-squared: R-squared: R-squared: R-squared: 0.26, 0.26 0.26 0.26 0.25 Adjusted RAdjusted RAdj. RAdjusted RAdjusted Rsquared: 0.21 squared: 0.21 squared: 0.21 squared: 0.21, squared: 0.21 F-statistic: F-statistic: F-stat. F-statistic: F-stat: 115.23 59.54116 39.33208 39.88 59.11 Durbin-Watson Durbin Watson Durbin-Watson Durbin-Watson Durbin-Watson stat: 1.794 stat: 1.776 stat.:1. 798 stat: 1.769 stat: 1.798 Variable

Table 8.3. Estimation results for transition economies; dependent variable: 'cit

'gdp 'bal 'bal*D1 Instruments:

Sample: 1990 2001, Total panel (unbalanced) observations: 105 0.88 0.89 0.94 (6.42) (6.54) (3.04) -0.41 -0.62 -1.66 (-2.12) (-2.53) (-0.24) 0.15 0.42 (1.42) (0.04) 'gdp and 'bal instrumented by 'gdp instrumented by 'bal 'gdp-1 'gdp-1 'cons-1 'def-2 R-sq: within = 0.60 R-sq: within = 0.61 R-sq: within = 0.3614

JACEK ROSTOWSKI

CHAPTER 9

HOW TO REFORM THE STABILITY AND GROWTH PACT

1. THE FUNDAMENTAL PROBLEM The most convincing justification for the Stability and Growth Pact (SGP) is the existence of a severe free rider problem in a monetary union (MU) with a number of large countries. Ever since it was discovered (probably by the Emperor Diocletian in the second century) that inflation can be used to reduce the government’s debt burden, a ‘fatal nexus’ has existed between government debt and governmental control over initially the mint, then the printing press and finally the central bank. Not only is it possible for a government that controls the creation of money to reduce its real debt burden by generating inflation, but this fact is known to money holders and affects their willingness to hold government created money. Negative effects on economic efficiency, savings, economic growth and even on so called seigniorage (government revenue from the creation of money) follow from the ‘option to inflate’ held by all governments. A number of mechanisms have therefore been invented to try and eliminate (or at the least severely limit) this option. They include fixing the parity of the state issued currency to one which is not so issued (the gold standard, currency boards, unilateral dollarization or euroization) and establishing an independent central bank. One problem has nevertheless persisted: how binding will such constraints on a government’s monetary behavior be when it actually faces the choice between debt default and breaking or rewriting such self imposed rules? Recent experience in Argentina suggests that ‘monetary rules’ may not be very binding in this situation, even if the final result in Argentina turned out to be both the breaking of the rules and government debt default. The result has been growing interest in ‘fiscal rules’ which are designed to prevent public debt from reaching a level at which it could become unserviceable, and might therefore present a temptation for government to inflate it away. Gordon Brown’s self-imposed so-called ‘golden rule’ is an example from Britain1, while the Polish constitutional limit on public debt to 60% of GDP is another. The problem of the link between fiscal and monetary prudence takes on an extra dimension in a MU which consists of a number of large, independent, fiscal

147 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 147-158. © 2006 Springer. Printed in the Netherlands.

148

JACEK ROSTOWSKI

jurisdictions. In the case of many small fiscal jurisdictions (countries or states) in the monetary union, a jurisdiction which borrows irresponsibly will become bankrupt before it affects the creditworthiness of the entire union (in the absence of a bailout promise from the others). This is the case of the United States. In the case of one very large and many small jurisdictions in the union, the small countries will be in the same situation as above, while the large jurisdiction will not be able to free ride on the reputation of the small ones, because of the large impact its indebtedness is likely to have on the exchange rate of the union currency. The eurozone seems to fit in the intermediate range, in which jurisdictions are small enough to be able to benefit from the credibility others give to the union currency, yet large enough for their bankruptcy to have large spillover effects on the union currency and other jurisdictions. Furthermore, the Germans have traditionally feared the following scenario: less monetarily and fiscally prudent member states (MS) would take advantage of the increased credibility and lower interest rates which a European MU would bring in order to increase their public debt to levels at which the monetary rules of the union would loose their credibility. In anticipation of the ensuing future inflation, interest rates would rise for all. In exporting German monetary credibility to all, the EMU would then end up by importing other members’ lack of credibility to Germany2. Nor is such a fear fanciful. Germany and other traditionally prudent countries such as Austria and the Netherlands are massively outvoted on the existing Governing Council of the ECB, and will be so even more after enlargement of EMU to the East. An additional problem is that the ECB does have some responsibility for the stability of MS’ banking systems3. Given the high levels of government debt held by banks in many EMU MS, in the case of default by an EMU member government it could be legitimately argued that this responsibility requires intervention to save the banking system of the country concerned. Given the extremely limited capital resources of the European System of Central Banks4 such intervention could only take the form of generating inflation5. There is also a further, deeper, problem. Whereas the ECB could allow small countries’ governments to default on their public debt, the consequences of allowing a major state to default and its banking system to collapse could be massively depressive for the whole union. In the absence of a MU, such a situation could be avoided by the affected country inflating the problem away. Within the eurozone it can only be avoided by the ECB generating inflation in the whole MU6. The special problem posed by the existence of a few large independent fiscal jurisdictions within a MU is highlighted by the absence of any federal level fiscal rules binding on individual states in the USA. This is partly the result of most states having state-level balanced budget rules. But it may also result from the fact that there are 50 states, with California (the largest) accounting for one eighth of the union’s GDP (and about one sixteenth of government expenditure in the USA), as compared to the 30% weight of Germany within the eurozone. Fiscal rules in a MU with a few large independent fiscal jurisdictions are therefore even more important for guaranteeing the long-term credibility of its monetary rules than they are in a unitary state.

HOW TO REFORM THE STABILITY AND GROWTH PACT

149

2. DEMAND SPILLOVERS There is little evidence for strong spill-over effects. Large econometric models have generated insignificant (or even negative) spill-over effects (Thygessen, 1998). An exception was the very large increase in the German deficit post-unification, which had perceptible positive spill-over effects on demand initially7. However, it is unclear that an excessive deficit provision helps a MU to deal with this problem, whereas a centralization on fiscal policy authority is unlikely to be politically acceptable. Thygessen (1998) suggests voluntary co-operation (as provided for by the Maastricht Treaty) as the least bad solution of the spill-over problem. 3. THE WEAKNESSES OF THE STABILITY AND GROWTH PACT 3.1. The excessive deficit procedure v. the Stability and Growth Pact The eurozone’s fiscal rules consist of two parts: (1) the excessive deficit provision and procedure and (2) the SGP proper. The excessive deficit and the procedures to deal with one are incorporated in the Treaty and its protocols. An excessive deficit exists is either of two conditions exists: (1) an EU member’s fiscal deficit exceeds the so-called reference value of 3% of GDP, or (2) its debt/GDP ratio exceeds 60%. All EU members are required to avoid excessive deficits (defined as above). EU members who belong to the eurozone can be punished on the basis of the so-called ‘excessive deficit procedure’ if they exceed the reference values for the fiscal deficit or public debt and are found to have done so by the Council of Ministers. Punishment can amount to fines of up to 0.5% of GDP yearly for persistent offenders. This is a large amount of money, current gross contributions to the EU being slightly over 1% of GDP, and the net contributions of large net contributors such as Germany and Italy being 0.2% of GDP. The SGP proper, on the other hand, consists of one political declaration by the European Council (of Heads of State and Government) and two regulations issued by the Council of Ministers8. The Pact does two main things: (1) it introduces the new obligation for states to achieve medium-term budgetary positions close to balance or in surplus; and (2) tightens up the procedures for identifying and for punishing excessive deficits among states which have adopted the euro. Thus, the excessive deficit procedure is thus quite strongly legally anchored, and allows for heavy fines on delinquent MS, while the SGP has a far less formal legal basis, and no procedure to punish a failure to comply with the objective of medium term balance. 3.2. Deficit vs. Debt aspects The effect of this greater legal and institutional weight given to the excessive deficit provision as compared to the SGP was compounded by the fact that the fiscal deficit requirement of the excessive deficit provision was treated with great seriousness, while its public debt/GDP requirement was largely ignored. As a result, primacy has been given to what is a short-term requirement (keeping the general government

150

JACEK ROSTOWSKI

deficit below 3% of GDP) compared to the medium-term requirement of achieving fiscal balance over the economic cycle and the long term requirement of limiting public debt to less than 60% of GDP. As we have seen, it is excessive public debt that constitutes the real threat to the credibility of a currency union’s monetary rules. Yet Italy, Belgium and Greece were allowed to join the eurozone with debt/GDP levels of over 100%. Furthermore, by focusing on the fiscal deficit aspect of the excessive deficit provision in the Maastricht criteria for eurozone accession, while keeping the SGP requirements legally weakly grounded and unsupported by effective sanctions, the EU created a situation in which the incentives for MS were highly asymmetric. States had a strong incentive to bring their deficits below 3% in order to qualify for eurozone membership. But once this had been achieved in 1998, they had little incentive to continue to improve their fiscal positions in the relatively good years of 1999-2001, so as to satisfy the SGP medium-term objective, given the absence of sanctions for such behavior. 3.3. The alleged ‘stupidity’ of the pact Many of the SGP’s problems stem from these ‘original sins’: 1. The fact that some MS failed to achieve medium-term fiscal balance in 1999-2001, meant that they had very little scope to allow automatic fiscal stabilizers to operate once conditions deteriorated in 2001-3, without coming up against the excessive deficit provision. 2. The fact that the fiscal deficit aspect of the excessive deficit provision was made paramount, instead of the public debt aspect, meant that a longer term approach, which would have allowed low debt countries to run larger deficits than those allowed by the Treaty when they ran into bad economic times had to be ruled out. Had either of these processes not occurred, we would not be talking today about a ‘stupid’ and ‘inflexible’ SGP. We have already discussed the reasons for the weakness of the medium-term goal of fiscal balance. There were also reasons for emphasizing deficit over debt in the excessive deficit provision: 1. The correct measurement of public debt is quite problematic. The measure chosen was gross public sector financial debt. Thus, a government with significant financial assets (e.g. in the form of loans to LDC borrowers in good standing) would be penalized. So would a state which undertook pension or health reforms which created ‘funded’ systems of finance for these needs, as that would transform so-called ‘implicit’ debt (which does not count under the Maastricht provision) into explicit debt which does9. 2. Had the excessive deficit provision successfully prevented deficits exceeding 3% of GDP, then it would likely have led to significant reductions in the debt/GDP ratio. The relationship between the deficit and the debt/GDP ratio is given by:

HOW TO REFORM THE STABILITY AND GROWTH PACT D/Y =

3.

gy * B/Y

151 (9.1)

Where D/Y is the fiscal deficit, gy is the growth rate of nominal GDP and B/Y is the steady state public debt/GDP ratio10. So that with the GDP deflator increasing at about 1.5% yearly and real GDP growth expected to be about 2% (giving nominal GDP growth of 3.5%), deficits on average of 2.5% would lead the debt/GDP ratio to converge on 71.4%. While this does not satisfy the excessive deficit provision, it is still a level which is unlikely to threaten government solvency11. However, the main reasons seem to have been political: a) There was no political will to exclude Italy, Belgium and Greece from the eurozone. b) There was an unwillingness to delay the launch of the EMU project until measurement problems had been resolved. c) The European Commission seems to have thought that stressing the fiscal deficit and the goal of medium term fiscal balance in the SGP, when combined with ‘multilateral convergence and surveillance procedure’12, could form the basis for the Commission gaining a key role in determining union-wide fiscal policy. Emphasis on the debt aspect of the excessive deficit provision would have left more room for policy discretion in the hands of MS, and thus undermined the hopes of the Commission.

3.4. Lack of enforceability However, the fundamental problem with the SGP is its lack of enforceability. Countries are not automatically declared in breach of the excessive deficit procedure. They can only be so declared by a vote of the Council of Finance Ministers, taken by qualified majority. In a situation in which several large and influential countries either have excessive deficits, or are close to having them, such a declaration is unobtainable, even though the country voted on is not allowed to participate in the voting13. In fact enforceability has not been uniformly weak. Small countries have until now on the whole respected both the excessive deficit proviso and the medium term goal of fiscal balance (the only serious exception has been Portugal). Table 9.1. Small is Beautiful (1997-2002)

Fiscal balance Growth Source: D.Gros [2004]

Big Three -2.06 1.93

Small Eight 0.13 3.95

Spain -1.29 3.55

UK 0.46 2.57

JACEK ROSTOWSKI

152

What we therefore have had so far is an ‘asymmetry of good behavior’. The fact that Ireland was actually censured in 2002 by the Council of Ministers for not increasing its surplus sufficiently suggests that we have also been witnessing the likelihood of an ‘asymmetry of enforcement’ (at least in the perceptions of the small countries themselves). However, with the refusal of the Council to punish France and Germany in November 2003, enforceability has presumably collapsed for all MS, both large and small. Whether this will lead to a sharp deterioration in fiscal performance among the small states is, however, unclear. Because of their size they can be allowed to default on their debt, with relatively limited impact on the eurozone economy as a whole. Knowing this, they may well fear sharply rising interest rates in the face of any large deterioration of their public debt position, and this may be sufficient in itself to dissuade them from lax fiscal policy. However, such responsible behavior by small MS would be only a limited consolation. It is after all the large states which can, by their fiscal irresponsibility, threaten the sustainability of the EMU’s monetary rules. Of course, France and Germany still have quite manageable public debt/GDP ratios, so there still is time to repair the breach in the SGP. The question is how that should best be done. 4. PROPOSALS TO IMPROVE THE PACT A number of proposals for the re-engineering of the Pact have been made. Four main kinds of mechanism have been suggested. 4.1. Expenditure rules Two main mechanisms have been proposed – ‘permanent balance’ and the so-called ‘golden rule’. Permanent balance involves forecasting future tax revenues and government expenditures and calculating their present value. As long as the two are equal, there is no fundamental shortfall of revenues, even if there is a large deficit at any particular moment (Buiter and Grafe, 2002). In particular, countries that expect to have high growth rates (and therefore much higher future revenues) can afford much higher deficits in the present. The idea is as well founded as its basic assumption – that the future can be forecast with confidence. The second mechanism within this approach is the UK Treasury’s golden rule, by which net spending on public investment is not included in the calculation of the excessive deficit (H.M.Treasury, 1998). The idea is that future generations, which will benefit from today’s public investment should contribute to it14. However, it is unclear why expenditure on physical infrastructure should count as public investment, while education does not. Also, if high investment is needed today because the current generation has under-invested (as is argued in the UK) then surely it is this generation (rather than future ones) that should pay for making up the backlog15. Finally, the real issue is the relative contribution to the net present value of the future stream of welfare generated by future GDP of public investment (either tax or debt financed) compared to the contribution of money left in private

HOW TO REFORM THE STABILITY AND GROWTH PACT

153

hands (and either consumed or invested). It is very possible that the money would be better used left in private hands. Public over-investment is also a possibility. 4.2. Allocation of deficit rights This could be done either by the EU itself, or by a pseudo market mechanism. It is very unlikely that MS would agree to give this power to the Commission, or that they could regularly agree on how to divide up the overall permitted deficit for the eurozone as a whole (which would be set jointly by the Commission and the Council of Ministers)16. Casella (2001) suggests that budget deficit permits should be tradable among MS (their total value would add up to a figure chosen each year at EU wide level). The key problem with this proposal is that the costs to the other members of the MU of a highly indebted large country increasing its deficit (and therefore its public debt) could be much greater than a slightly indebted small country doing so17. 4.3. An Independent Fiscal Authority Wyplosz (2002) has suggested the creation of independent Fiscal Policy Committees at national level in all eurozone MS. These would be to the macro-policy fiscal stance what independent central banks are to monetary policy. They would base their decision on the fiscal stance each year on the need to ensure long-term sustainability while minimizing short-term GDP and price fluctuations. Even if politicians agreed to the de-politicization of the fiscal stance, it is unclear how ministers of finance could vary expenditure and revenue so as to achieve the stance set by the FPC efficiently. One possibility would be for most of the short-term changes to be achieved through tax rate changes, with expenditure variations being much slower. A possible difficulty in this case might be a relatively low level of response of aggregate demand to what would be known to be transitory tax changes (European Commission, 2002). However, the strength of the proposal is that it sharply reduces the power of national politicians over fiscal policy, while at the same time establishing a framework which is designed to ensure long-term fiscal sustainability. 4.4. Making Debt Sustainability Central to the SGP Pisani-Ferry (2002) suggests allowing countries whose public debt/GDP ratio is less than 50% to opt for a ‘debt sustainability pact’ instead of the excessive deficit procedure. Countries which did so would be required to provide estimates of the future impact of current budgetary commitments (such as the pension debt implicit in PAYG systems). Like the previous one, the proposal puts the stress where it should be – on sustainability. Its weaknesses are enforceability and the fact that implicit debt calculations are very sensitive to assumptions.

154

JACEK ROSTOWSKI

5. ENTRENCHED NATIONAL GUARANTEES OF DEBT SUSTAINABILITY The most important problems with fiscal rules in the EMU is their enforceability and their legitimacy. Legitimacy itself requires: 1. That the rules be seen to be necessary; 2. That the rules be enforced in an impartial and equal manner on all states in the eurozone. As we have seen, rules aimed at maintaining the sustainability of public finances are relatively easy to justify, while the need to avoid spill-over effects on aggregate demand is far less clear. As long as the imposition of sanctions for breaches of the SGP remains a political decision in the hands of national governments, large MS will be unlikely to be punished. Yet it is precisely these states that most need to be controlled if the monetary rules of the eurozone are to be protected. There are two possible solutions to this conundrum: the first is to give far greater power in the fiscal sphere to the European Commission, which would be allowed both to issue warnings and to impose fines under the excessive deficit procedure. I find such a degree of centralization uncongenial and unlikely to be politically acceptable to MS18. The alternative is to require eurozone states to enact domestic fiscal rules, which would be entrenched (i.e. would require supermajorities to rescind) and which if rescinded, would lead to the countries concerned having to quit the eurozone. An interesting example of such rules is the Polish constitutional limit on the public debt/GDP ratio, which may not exceed 60%19. It has the advantage of being both relevant (applies to debt sustainability) and enforceable. Countries that have adopted the euro could be required to enact a provision of this kind, so that budgets that caused the public debt/GDP ratio to exceed the limit would simply not have legal force20. The rules could be more detailed, setting out what was required of budgets in the case when the ratio exceeded lower thresholds of 50% and 55%, as is done in the Polish “Law on Public Finance”21. Domestic fiscal rules cannot, by definition, be violated (a budget which violates such rules is simply not legal, and could not be enforced in domestic law). They can, however, be rescinded by sufficient super-majorities. However, under my proposal, countries which rescinded these rules (and thus break the conditions of eurozone membership) would automatically be required to leave the eurozone22. This is presumably what would happen under present rules if a country broke the Treaty’s requirement that its central bank be independent23. Thus, by combining domestically entrenched fiscal rules (which simply cannot be legally broken) with the external anchor of their existence being a requirement for continued eurozone membership, we get fiscal rules that exhibit a very high degree of enforcement and yet do not centralize discretionary fiscal power at MU level. Given the drastic nature of expulsion from the MU as a sanction for breaking the public debt/GDP limit, it would be useful (as in the Polish case) to have lower thresholds in the domestic fiscal rules, the passing of which would require remedial action before the ultimate limit is reached. At present, Poland’s Law on Public Finance forbids any increase in the budget deficit relative to budget revenues once the ratio of public debt/GDP exceeds 50%, and any increase in the debt/GDP ratio

HOW TO REFORM THE STABILITY AND GROWTH PACT

155

once the 55% threshold has been passed. However, the weakness of these regulations lies in the fact that, if the deficit is higher than planned in year t, the debt/GDP ratio can be certified as higher than 50% only in March of the subsequent year (t+1). As a result, the ban on increasing the deficit (compared to that of year t) then applies only in year t+2. Thus, the Polish Law needs to be changed , so as to prevent a government ‘jumping’ the lower thresholds on the growth of public debt by increasing the deficit very fast (as happened during 2003-2004), and leaving its successor with the need to rein debt in drastically, since it is right up against the constitutional 60% limit. Such a reform in the Polish case might consist in a requirement that deficits not exceed 3% if an even lower threshold of public debt/GDP (say 45%) is passed. Leaving aside the requirements of domestic law, within a reformed EMU, MS that violate a suitably chosen lower debt threshold could also loose their right to vote on the General Council of the ECB and in the ‘eurogroup’ of finance ministers (which decides on fiscal policy co-ordination). This would prevent such countries’ representatives from voting in favor of inflationary policies that would help reduce their real debt burdens24. Expulsion from the eurozone would not necessarily mean that a country would be obliged to reintroduce its own national currency to replace the euro. Countries might decide to continue using the euro, but their central banks could no longer participate in its creation, in the setting of interest rates for the euro area, in seigniorage revenue from euro creation25 or in any implicit ‘lender of last resort’ liquidity support which their banking systems might expect. They would then effectively be ‘unilaterally euroized’. At the very least an extremely strong signal will have been sent to the markets that the country is not in the same ‘risk league’ as eurozone members, and that the risk of government default depends exclusively on the condition of its own public finances. Re-admission to the eurozone after expulsion would be made conditional on reintroduction of the domestic fiscal rules required by the Treaty, and on bringing the public debt/GDP ratio down below the Treaty’s lower debt threshold. There are, however, serious problems in implementing the proposal. The main one is that of ‘unequal starting points’. Italy and Greece (and to a lesser degree Belgium) have public debts that would be unacceptable if they obtained in the whole of the MU. We either have to accept that some countries are allowed to maintain higher public debts than others are allowed, or these debts must be reduced. A possible solution would be to give these countries a number of years derogation from the debt/GDP limit, during which time they would have to bring their ratios into line, or be automatically excluded from the union. The second problem is implicit pension debt. Pension reforms which move a country to a system of ‘funded’ pensions transform implicit pension debt into explicit government debt to the pension funds. Nearly every single country implementing such a reform would find that its public debt/GDP ratio would exceed 60% at the end of the transition26. A 60% limit will thus discourage the implementation of reforms, so a higher limit needs to be agreed for countries with funded pension systems.

156

JACEK ROSTOWSKI

But what should that limit be? And would a limit of say 200% or 250% of GDP (which would be required for many countries after pension reform) lead to a sustainable debt? The two kinds of debt (implicit and explicit) are not, after all, fully comparable. Explicit debt is ‘harder’. Governments cannot renege on it as easily as they might be able to reduce expected (or current) pensions in case of severe fiscal stress. 6. CONCLUSION: EMU IS A “PREMATURE MONETARY UNION” The EMU is a ‘premature monetary union’. It is premature in its overall structure (i.e. the absence of an effective fiscal constitution – which it vitally needs), and also in the admission of countries that should not have been admitted given that their debt/GDP ratios were at levels that would not be safe for the union as a whole (Greece, Italy and maybe Belgium). It is also premature in the sense that the rules of admission to the union are not well designed, particularly as regards the choice of the exchange rate at which currencies will be converted into euros. One question flowing from this analysis is whether EMU can be repaired ‘on the go’ (is it, figuratively, a ship or a plane?). Second, given the political difficulty of enacting the reforms I have suggested, the usefulness of such proposals (and many of those made by other authors) may be doubted. However, EMU may not be the last MU to be created. There has been talk of an East Asian Monetary Union. So analyzing the errors that have been made in setting up the EMU need not be a pointless task, even if EMU itself proves unreformable.

NOTES 1

By this rule fiscal deficits should be balanced over the economic cycle except to the extent to which they pay for net public investment. 2 This may be part of the reasoning behind German opposition to unilateral euroization, although this is misplaced, as a unilaterally euroized country has no right to ECB support for its banking system and would be unable to generate euro inflation itself (see Bratkowski and Rostowski, 2002). 3 By Art.3 of the ECB Statute, the European System of Central Banks (ESCB) to which the ECB is central, has the responsibility of promoting the smooth functioning of payments systems. 4 The ECB and the National Central Banks of the eurozone MS. 5 This is why Ivo and Lemmen (1999) propose that government debt within a MU should not be treated as having zero risk as it is under present Bank for International Settlements (BIS) guidelines. Such a move is justified by the formal unavailability of the inflation option within the eurozone. Moreover, setting limits on single government debt exposure by banks would encourage diversification of banking system assets, making banks less exposed to default by their home country government (although partly at the cost of making them more exposed to default by other eurozone governments). 6 I exclude the possibility of the ECB providing the affected banking system with liquidity, since the problem we are discussing is one of the solvency of the government concerned and therefore of the banking system of the country as well. Such a problem cannot be resolved by mere injections of liquidity. 7 The subsequent increase in interest rates and ensuing fall in investment demand was a ‘pecuniary externality’, i.e. one in which markets transmit the right signals from one country to another. European (and indeed world-wide) interest rates do (and should) increase in response to a large increase in demand

HOW TO REFORM THE STABILITY AND GROWTH PACT

157

for financing by a major government. In the absence of a MU a larger part of the adjustment would occur through an appreciation of the currency of the state concerned. In the presence of the MU more of it has to occur through an increase in interest rates throughout the union. However, such an increase is preferable to some other form of rationing, and may well be preferable to not allowing the deficit to occur in the first place. 8 The first made at the Amsterdam summit of 1997, the latter a few weeks later. 9 One valid reason for leniency to Belgium in spite of its high debt, was that it has high levels of funded provision for both pensions and health care. 10 The derivation of the formula is straightforward. The yearly change in the public debt equals the deficit of the public sector, so that ˜B/˜t (the change in public debt) = D. Now B = (B/Y)*Y. If we take the time derivative of B we get: ˜B/˜t = (B/Y)˜Y/˜t + Y˜(B/Y)/˜t. After setting D = ˜B/˜t we get: D = (B/Y)˜Y/˜t + Y˜(B/Y)/˜t. Dividing by Y, we get: D/Y = (B/Y)(˜Y/˜t)/Y + ˜(B/Y)/˜t =

gY(B/Y) +

˜(B/Y)/˜t. Now in the steady state ˜(B/Y)/˜t = 0, so we are left with: D/Y = gY(B/Y). 11 The Maastricht Treaty limit on B/Y < 0.6 together with the deficit limit on D/Y < 0.03, therefore implies a nominal growth rate of 5%. Even with inflation at (rather than under) the 2% p.a. ceiling set by the ECB, this implies a real growth rate of 3% p.a., which is improbably high. 12 The Commission develops Union-wide ‘broad economic guidelines’ annually, and MS are then required to develop ‘stability programs’ (for eurozone members) and ‘convergence programs’ (for noneurozone members) which must be consistent with the BEGs. 13 In November 2003 France, Germany and Portugal were projecting excessive deficits for the fourth year running, and Italy was close to an excessive deficit. 14 Which is why amortization of past investment should be included in any excessive deficit. 15 This point was made by Dr. Bini-Smaghi at the Aspen European Dialogue conference, Venice 26-27 February 2004. 16 The proposal was originally made by Strauss-Khan in 1999 as a way of setting the maximum permitted deficit for the whole eurozone at 3% of GDP (Buti, Eijfinger and Franco, 2003). Since different MS would often be at different stages of the business cycle, this would allow many to have deficits far in excess of 3% and would sharply increase the average deficit of the whole eurozone. 17 A debt-weighted premium could be added bureaucratically to the price. Alternatively, one could rely on the low deficit members of the union charging a higher price to a high risk purchaser of deficit rights so as to protect themselves from the spill-over effects. 18 Buti et al. (2003) have the same view. 19 By being part of the constitution, it requires a two thirds majority in the lower house (Sejm) of the Polish parliament, together with a simple majority in the Senate, to overthrow it. 20 This could present a problem in the case of the UK which does not have a written constitution or entrenched legislation. 21 This is an ordinary, un-entrenched, provision. 22 The new treaty (which would be needed for the implementation of this approach) would state categorically that countries which rescind the required legislation, automatically leave the eurozone. 23 The Governor of its Central Bank would presumably be excluded from the ECB Council under such circumstances, and if the central bank began to credit the government or “parastatal agencies” the country would presumably be asked to leave the ESCB and the eurozone. 24 The present deposits and fines for eurozone members running an excessive deficit do not seem to be incentive compatible, as they make fulfillment of the Treaty requirements even harder. 25 In the case of countries which are net contributors to the ECB (such as Germany) this could actually improve their financial position if the demand for euro money base growth was small or negative, and the net contribution to the ECB (which is the difference between earnings from assets backing the monetary liabilities of the national central bank concerned and the country’s share in total eurozone seigniorage based equally on its share in eurozone GDP and population) was large. 26 With the possible exceptions of the UK and Ireland.

158

JACEK ROSTOWSKI REFERENCES

Bratkowski & Rostowski (2002). The EU attitude to unilateral euroization: misunderstandings, real concerns and sub-optimal admission criteria. Economics of Transition, 10 (2), 445-468. Buiter, W. & Grafe, C. (2002). Patching up the pact – some suggestions for enhancing fiscal sustainability and macroeconomic stability in an enlarged Euiropean union. CEPR Discussion Paper, 3496. Buti, M., Eijfinger, S. & Franco, D. (2003). Revisiting the stability and growth pact: grand design or internal adjustment? European Economy, January, from http://europa.eu.int/comm/economy_finance Casella, A. (2001). Tradeable deficit permits. The Stability and Growth Pact – The Architecture of Fiscal Policy in EMU, Palgrave, Basingstoke. European Commission (2002). Public finances in EMU. European Economy, June, from http://europa.eu.int/comm/economy_finance/publications/european_economy/2002/ee402en.pdf Gros, D. (2004). Learning from each other: competing for structural reforms in Europe. Paper presented to the Aspen European Dialogue Conference at Venice, February 27-29. H.M. Treasury, (1998). The code for fiscal stability. London. Ivo, A. & Lemmen, J. (1999). The vulnerability of banks to government default risk in EMU. Financial Markets Group Special Paper, 115, July. Pisani-Ferry, J. (2002). Fiscal discipline and policy co-ordination in the Eurozone: assessment and proposals. Paper prepared for the Group of Economic Analysis of the European Commission, Brussels. Thygessen, N. (1998). Fiscal institutions in EMU and the stability pact. Social Challenges of Economic and Monetary Union eds. P.Pochet and B.Vanhercke, European Interuniversity Press. Wyplosz, C. (2002). Fiscal policy: rules or institutions? Paper prepared for the Group of Economic Analysis of the European Commission, Brussels.

NIKOLAI ZOUBANOV

CHAPTER 10

UNEVEN GROWTH IN A MONETARY UNION

1. INTRODUCTION The issues raised by the accession to EMU of new, fast growing countries (FGC) reflect a much larger problem: that of uneven growth between regions of similar size within a monetary union (MU)1. One can easily imagine a situation in which one region of EMU (say France and Spain) grows fast, while another (say Germany and Italy) of about the same size grows very slowly. What will be the effects of such a situation on the slow-growing region? Although there are other factors influencing inflation and the external exchange rate in a MU, we focus on the role of the HarrodBalassa- Samuelson (HBS) effect in FGC for slow-growing countries (SGC) and the whole union. This chapter is organized as follows: Section 10.2 provides a brief introduction to the problem. Section 10.3 presents a mathematical model of the HBS effect in one country (not in a MU), and derives implications for the exchange rate and for inflation in FGC outside a MU, assuming a tight monetary policy stance. Section 10.4 augments the model with the case of two countries in a MU with uneven growth rates (faster- and slower-growing), and derives the implications of productivity growth in the FGC for the SGC. An outline of policy implications following from theoretical discussion is offered in Section 10.5. 2. UNEVEN PRODUCTIVITY GROWTH AND INFLATION DIFFERENTIALS If the exchange rate is fixed, the FGC will experience higher inflation than will the SGC. To illustrate this, consider the following model (see De Grauwe, 1992). There are two countries in a MU, A and B, producing tradable and non-tradable goods. Tradable goods are easy to export, so their prices tend to be equal in both countries. Non-tradable goods, though, have such large trade costs that they never enter international trade, hence their prices may vary from country to country. It is further assumed that the share of non-tradable goods in total output is a (so that (1-a) is the share of tradables), and these shares are equal in the two countries, and the wage level is the same in both sectors (although not necessarily the same in both countries). The inflation rate (ʌ) in A and B is defined as follows: 159 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 159-182. © 2006 Springer. Printed in the Netherlands.

NIKOLAI ZOUBANOV

160

SA SB

(1  D )S TA  DS NA (1  D )S TB  DS NB

(10.1)

where subscripts T and N stand for tradables and non-tradables, respectively. Because of the existence of a single currency, inflation rates in tradables are equal across countries. At the same time, inflation in each sector should be equal to differences between the growth rates of wages and of labor productivity in respective sectors, that is:

S TA wˆ A  qˆTA S NA wˆ A  qˆ NA S TB wˆ B  qˆTB S NB wˆ B  qˆ NB

(10.2)

where xˆ d log x is the rate of change in relative terms. We can now derive nontradable inflation as a function of tradable inflation and productivity in both sectors:

S NA S TA  qˆTA  qˆ NA S NB S TB  qˆTB  qˆ NB

(10.3)

By substituting these back to equation 10.1, we get:

SA SB

(1  D )S TA  D (S TA  qTA  q NA ) (1  D )S TB  D (S TB  qTB  q NB )

(10.4)

which implies

S A  S B D (qTA  qTB )  D (q NA  q NB )

(10.5)

If we assume that there is equal productivity growth in non-tradables in both countries, we obtain the following equation:

S A S B

D (qTA  qTB )

(10.6)

which implies that if A’s labor productivity in tradable sector grows faster than B’s, A will experience higher inflation than will B. So, uneven productivity growth implies uneven inflation in a MU, which is unavoidable, unless the growth is stopped.

UNEVEN GROWTH IN A MONETARY UNION

161

Although this model is a convenient way to start exploring the impact of uneven growth on the economy of MU, it does not address a number of important issues. In particular, what happens to employment and income in the member countries? How can inflation be managed by means of monetary policy? What are the opportunities and challenges for the SGC being in a MU with an FGC? These and some other questions are the ones we will try to answer by exploring a model of the HBS effect in detail. 3. THE HBS EFFECT AND ITS IMPLICATIONS FOR THE CASE OF ONE COUNTRY In this section, we explore the implications of the HBS effect for the price level, output, domestic demand and export potential in the case of one country, and also link it with the monetary policy stance so as to show its implications for the exchange rate. It will be shown that productivity growth in the tradable sector leads to appreciation of the country’s real exchange rate, but that monetary policy can be used to reach a trade-off between nominal exchange rate appreciation and inflation. Besides, in no case will a country experiencing productivity growth lose its competitive advantage in real terms. It has a wide choice of exchange rate/inflation tradeoffs which do not hurt its position on the world market. 3.1. Assumptions Consider a country S with a small open economy that takes the price of imports and exports as given on the world market. There are two sectors in the economy: tradables (goods available for trade worldwide with reasonable costs of shipping and other trade barriers), and non-tradables (goods so costly to trade that they never enter the international trade). The relative unit composite price of non-tradables in terms of tradables is set equal to pS, with an initial value of 1. There are two factors used in production: labor and capital. While S is able to freely import capital from abroad at a fixed interest rate r, the domestic supply of labor is limited and equal to Ls=LTS+LNS, where subscripts TS and NS stand for the tradable and non-tradable sectors in S, respectively. All factor markets in S are perfectly competitive, which implies that wages are equal across the whole economy. Some other assumptions also apply: 1. Output in both sectors is determined by the Cobb-Douglas production function with two factors, labor and capital, and constant returns to scale:

2.

D TS

YTS

ATS K TS

LTS

YNS

pS ANS K NS

1D TS

D NS

LNS

1D NS

(10.7)

Consumption is regulated by a constant-elasticity-of-substitution (CES) utility function:

NIKOLAI ZOUBANOV

162

TS

u S (C )

T S 1 T S 1 T 1 1 ª 1 º S TS TS T T «J S S ˜ CTS  (1  J S ) ˜ C NSS » ¬« ¼»

(10.8)

subject to the budget constraint: Zs = CTS + psCNS = wsLs + rQs

(10.9)

where ws – equilibrium wage level in S, Qs = domestic capital plus current account balance, CTS, CNS stand for consumption of tradables and non-tradables, respectively, 0<ȖS<1 is the share of tradables in total consumption,

TS

3.

§C d log¨¨ TS © C NS d log pS

· ¸¸ ¹ is the elasticity of substitution between tradables and non-

tradables, meaning that with a 1% increase in pS consumption of tradables grows by șS% relative to that of non-tradables. The nominal exchange rate between S’s currency and other currencies in the world is defined by

H S ,i

4.

5.

PTi PTS

(10.10)

where PTS,Ti are the nominal prices of tradables in countries S and i. Note that we take into account only the prices of tradables, since for the nominal exchange rate (which is formed as a result of interaction of supply and demand for currencies required to buy goods on international markets) the prices of non-tradables should not matter. The real exchange rate is, according to the theory of purchasing power parity, initially equal to one for all the countries, which implies that the same bundle of consumption goods costs the same in all countries. S experiences productivity growth in its tradable sector, ÂTS=dlogATS>0, and no productivity growth in its non-tradable sector. Also, for simplicity, it is assumed that the outer world has zero productivity growth. The assumption of no productivity growth in S’s non-tradable sector is not crucial to the result, yet very helpful for deriving the algebra. Very similar calculations can be performed for the general case of non-zero productivity growth, but with considerably more technical difficulty.

UNEVEN GROWTH IN A MONETARY UNION

163

3.2. The model We approached the derivation of the HBS effect when we modeled the impact of uneven growth on inflation differentials between two countries. The model we considered there was too simple to catch any other effect than the inflation differentials, and now we have to do some more careful modeling to show other implications of productivity growth. We start with the case of one country to show the basic intuition, and then will proceed to the case of two countries to see the ‘spillover’ effects of productivity growth from one country within a MU to another. Throughout this chapter we use the modeling framework of Obstfeld and Rogoff (1996), with some extensions and minor alterations. Define the output per employed unit of labor as y=Y/L. Then, given a CRS Cobb-Douglas production function (equation 10.7),

y NS ,TS

( pS ) ANS ,TS ˜ k NS ,TS

D NS ,TS

(10.11)

where k=K/L. By assumption of perfectly competitive factor markets and constant returns to scale,

yTS

ATS ˜ kTS

D TS

rkTS  wS and y NS

pS ANS ˜ k NS

D NS

rk NS  wS (10.12)

Log-differentiate equations (10.12) to get

yˆTS , NS

rkTS , NS rkTS , NS  wS

kˆTS , NS 

wS rkTS , NS  wS

wˆ S

(10.13)

which is equivalent to

AˆTS  D TS kˆTS pˆ S  Aˆ NS  D NS kˆNS

D TS kˆTS  (1  D TS ) wS

and

D NS kˆNS  (1  D NS ) wS

(10.14)

where we made use of the fact that the shares of capital and labor in the total output are equal to Į and (1-Į), respectively. Define 1-Į=µ. Then

Aˆ TS which gives us

P TS wˆ S and pˆ S  Aˆ NS

P NS wˆ S

(10.15)

NIKOLAI ZOUBANOV

164

pˆ S

P NS ˆ ATS  Aˆ NS PTS

(10.16)

But remembering that we assumed no growth in S’s non-tradable sector, equation 10.16 simplifies to

P NS ˆ ATS PTS

pˆ S

(10.17)

This can be interpreted as follows: the elasticity of ps with respect to ATS is

P NS PTS

. That is 1% increase in tradable sector productivity causes

P NS PTS

% increase in

relative price of non-tradables. By maximizing the utility function (equation 10.8), we get the following demand schedules:

CTS

T

J S ZS , J S  (1  J S ) pS 1T

C NS

S

pS S (1  J S ) Z S J S  (1  J S ) pS 1T S

(10.18)

By substituting equations 10.18 back into equation 10.8 and scaling it down to unity, we get the overall price level (defined as minimum expenditure such that C=1),

PS

>J

S

 (1  J S ) pS

1 1T S 1T S

@

(10.19)

Log-differentiate equation 10.19 to obtain

PˆS

(1  T S )(1  J S ) pˆ S 1 1  T S J S  (1  J S ) p S 1T S

>

@

(10.20)

Remembering that we have set up ps=1 initially, equation 10.20 simplifies to

PˆS

(1  J S ) pˆ S

(10.21)

This result suggests that the real aggregate price level in S increases by (1-Ȗs)% with each 1% increase in ps, or, equivalently, by

P NS PTS

(1-Ȗs)% with each 1%

UNEVEN GROWTH IN A MONETARY UNION

165

increase in tradable sector productivity. So, we have derived the HBS effect in its initial formulation: wealthier countries tend to have higher price levels. With output growing in S due to increased productivity, its wealth increases, but so does its price level. Now comes the implication for the real exchange rate: assuming that there is no productivity growth in the reference country i, S’s real exchange rate related to i’s currency appreciates, as S has become a more ‘expensive’ country with an increase in productivity there. 3.3. Further implications Following the theoretical model of the HBS effect, we can also derive some implications of that effect in S for its output per worker, total output, income growth, domestic demand and allocation of labor in the economy, and finally link the HBS effect with inflation and the exchange rate through monetary policy. 3.3.1. Changes in output per-worker Recall that

yTS

ATS kTS

D TS

rkTS  wS (equation 10.12). Differentiate with

respect to kTS to get:

D TS ATS kTS D

TS

1

r

(10.22)

Then log-differentiate equation 10.22 to get:

aˆTS  AˆTS  (D TS  1)kˆTS

rˆ

(10.23)

Because r and Į are constant, the above reduces to

kˆTS

AˆTS

PTS

(10.24)

Combining equations 10.12 and 10.24, we obtain:

yˆTS

kˆTS

(10.25)

which means that with a 1% productivity growth in tradables, per-worker output in this sector grows by NS, we get:

1

PTS

%. If we redo these calculations for the non-tradable good,

NIKOLAI ZOUBANOV

166

yˆ NS

kˆNS

pˆ s  Aˆ NS

pˆ s

P NS

P NS

(10.26)

which is equal to the growth of output per worker in TS, if we apply equation 10.16. This result is paradoxical, but only at the first sight: since the wage and interest rate are the same in all sectors of the economy, labor and capital are employed up to the point where their marginal products are equal for all sectors. Actually, this is one of the implications of the HBS effect: the relative price of non-tradables makes up for the gap in real productivities between tradables and non-tradables. 3.3.2. Changes in wages and total income Recall from previous derivations that

yˆTS

AˆTS

PTS ˜ wˆ S ,

or

wˆ S

Aˆ TS

P TS

. So,

wˆ S . This says that the increase in wages is fully compensated by the labor

productivity increase, then labor still receives its marginal product and the nominal price of tradables (if we were to introduce one at this point) stays unchanged, ceteris paribus. This result goes to show that any other country trading with S will not have inflationary pressure from importing TS, because its price does not change due to productivity increase. Recall the budget constraint (equation 10.9), Zs = wsLs + rQs. Clearly, with a 1% increase in ws, Zs grows by grows by

M % PTS

M

wS LS ˆS %. Since w wS LS  rQS

with a 1% increase in productivity, or by

AˆTS

pˆ S

PTS P NS M % with a P NS

, Zs 1%

increase in ps. With income increasing, demand should also increase, but its increase need not necessarily match the output increase, so that we may have a change in S’s export potential, which we will not be able to calculate before we explore how domestic demand reacts to the increase in productivity. 3.3.3. Changes in domestic demand Log-differentiate the optimal demand schedules (10.18) to get

CˆTS ZˆS  (1J S )(1TS ) pˆ S and Cˆ NS

Zˆ S  >T S  (1  J S )(1  T S )@ pˆ S (10.27)

UNEVEN GROWTH IN A MONETARY UNION Because

M pˆ S , P NS

Zˆ S

M  (1  J S )(1  T S ) % P NS

167

the net increase in domestic demand for tradables is per

each

1%

increase

in

ps,

or

º P ª M  (1  J S )(1  T S )» ˜ NS % per each 1% increase in ATS. Analogously, the « ¼ P TS ¬ P TS net increase in demand for non-tradables is

M  >T S  (1  J S )(1  T S )@ % and P NS

ª M º P  >T S  (1  J S )(1  T S )@» ˜ NS % per each 1% increase in ps and ATS, « ¬ P TS ¼ P TS respectively. Note that, reasonably assuming T S ! 0 , consumption of non-tradables rises more slowly than that of tradables, which implies that in growing economies the share of non-tradables in consumption should gradually fall, thus lowering the impact of the HBS effect on the overall price level. While overproduced tradables may be exported, non-tradables must be consumed within the country. Since per-worker productivity does not necessarily have to grow by the same amount as demand, there is a possibility of reallocation of labor between tradable and non-tradable sectors, which is discussed next. 3.3.4. Reallocation of labor Recall that the physical

YNS

ANS ˜ k NS

D NS

Lˆ NS

output

of

non-tradables

is

equal

to

˜ LNS . Log-differentiate this, remembering that ÂNS=0, to get: YˆNS  D NS kˆNS

Cˆ NS  D NS kˆNS

(10.28)

since non-tradables do not enter international trade. We already know what ƘNS and

kˆ NS are (equations 10.26 and 10.27), so just by bringing them together we get: ˆ

LˆNS

>M  PNS ˜ TS  (1J S )(1TS )  (1 PNS )@˜ ATS ,

(10.29)

or LˆNS

>M  PNS ˜ TS (1J S )(1TS ) (1 PNS)@˜ pˆS

(10.29a)

PTS

PNS

NIKOLAI ZOUBANOV

168

We can see from this that – unless S has negative financial wealth so that its share of labor in output is above 1 – the higher T S , the lower is the growth of employment in the non-tradable sector. Indeed, if non-tradables are well-substituted for by tradables ( T S approaching 1 from below), then there is no need to produce more non-tradables, because the actual demand for them will not grow even though income will (if T S =1 and ij<1 employment in non-tradable sector should actually fall). On the other hand, if non-tradables are not easily substituted for by tradables, there will be a higher growth of employment in that sector to meet the growing demand. Theoretically, there need not be an increase in unemployment because of labor reallocation: production of non-tradables will exactly meet the demand for them, and every extra unit of tradables produced will be sold on the world market. However, because of structural rigidities and imperfect competition on the world market, a country experiencing productivity improvement may find it difficult to sell all of its extra output of tradables abroad and then unemployment may rise. Analysis of this situation is out of this topic, but one can find casual evidence of this in Gordon (1998). 3.3.5. Changes in total output In this paragraph we continue to assume that labor reallocation does not cause unemployment. Before the productivity shock, the quantity of labor employed in the non-tradable sector was

J S ( wS LS  rQS ) P NS wS

' , and therefore LS-ǻ units of

Lˆ NS %

labor were employed in tradables. If with a 1% increase in ps there is

increase in the labor employed in non-tradables, then the tradable sector is going to have



LS  ' ˜

100  Lˆ NS 100

units

employed,

thus

growing

by

' ˜ Lˆ NS 100 ˜100 Q % 2. Then the tradable output grows at LS  '

YˆTS

which is

1

P NS

yˆ TS  LˆTS

pˆ S

P NS

 Qpˆ S

(10.30)

 Q % per each 1% increase in ps. The growth of total output is a

weighted average of the growth rates of tradable and non-tradable outputs3.

UNEVEN GROWTH IN A MONETARY UNION

169

3.3.6. Changes in export potential Given that consumption of tradables grows by tradable output grows by export potential is

1M

P NS

1

P NS

M  (1  J S )(1  T S ) %, and that P NS

 Q % (equation 10.30), the net increase in S’s

 (1  J S )(1  T S )  Q % per 1% increase in ps. This is

very likely to be positive, suggesting that a country experiencing productivity growth tends to have its current account balance improved through exports. The extent to which it can do so depends on the economy’s structural parameters and consumer preferences. A high share of labor employed in the zero-growing nontradable sector dampens the effect of productivity growth in tradables. A high share of labor income in GDP creates higher internal demand for tradables, because it is the workers who benefit from the improvement in productivity first. Low substitutability between tradables and non-tradables reduces the effective internal demand for tradables thus increasing export potential. On the other hand, if nontradables are easily substituted by tradables, there is little room for a substantial increase in export potential, because much of the increase in output would be ‘eaten up’ by domestic consumers. 3.3.7. Changes in social welfare and consumer behavior Social welfare improves with the increase in consumption, provided the share of labor income in GDP is high enough4, non-tradables are substitutable for tradables and there are no significant structural rigidities on the factor markets. Furthermore, along with the obvious result that output grows with improvements in productivity, we get a less obvious result that S is very likely to get richer in terms of purchasing power: while its aggregate real price level rises by (1-ȖS)% with each 1% increase in pS, its income goes up by

M % P NS

for the same increase in pS, not to mention an

increase in its current account surplus due to more exports. One implication of this is growing demand for imports, which is an important insight for the next part of the discussion – the case of two countries in a MU. 3.4. A monetary policy link Notice that all the derivations above do not depend on nominal prices (in other words, they are correct for any numbers), since they were calculated for relative prices. But now we can link the HBS effect with monetary policy and the nominal exchange rate. Although the HBS effect is modeled using relative prices, we can

NIKOLAI ZOUBANOV

170

introduce money in this model if we assume a certain monetary policy stance. Consider two possible different cases of a ‘neutral’ monetary policy: (i) when the money stock grows proportionally to the GNP (that is when M/Y=const); (ii) when the money stock is constant irrespective of output changes. 3.4.1. The case when M/Y=const Recalling that MV=PY, M/Y=const implies that the aggregate nominal price level, Pn, stays constant unless there are changes in the money velocity. Since the relative price of non-tradables rises, in order to guarantee that the average price level, Pn, remains constant, its base, that is the nominal price of tradables, should fall. But in that case, because of competitive forces, the country’s nominal exchange rate must appreciate, thus bringing the price of its tradables back to the international level. Formally, the nominal price level is equal to the real price level multiplied by some number Ș, later called the numeraire, for which it is convenient to use the nominal price of tradables:

Pn

>

1 1T S

@

(10.31)

Kˆ  (1  J S ) pˆ s

(10.32)

K ˜ PS K ˜ J S  (1  J S ) pS 1T

S

Log-differentiating (10.31) we get:

Pˆn

Kˆ  PˆS

Then, to keep the above constant, we need

Kˆ (1  J S ) pˆ s ,

that is the

numeraire should fall by (1-Ȗs)% with each 1% increase in ps, or, equivalently, by

P NS PTS

(1-Ȗs)% with each 1% increase in tradable sector productivity.

Recalling the formula for the nominal exchange rate between countries S and i (equation 10.10) and log-differentiating it, we obtain:

HˆS ,i

PˆTi  PˆTS

PˆTi  Kˆ

(10.33)

Since we have assumed that there is no productivity growth in the outer world, PTi should not change, and then we have the nominal exchange rate appreciating by (1-Ȗs)% with each 1% increase in ps. Perhaps it is now needless to say that S does not lose its competitiveness because its exchange rate appreciation is exactly matched by the fall in its domestic price of tradables5. 3.4.2. The case when M=const Constant money supply implies a reduction in the aggregate price level, and the price of tradables falling even more than in the previous case. The exchange rate

UNEVEN GROWTH IN A MONETARY UNION

171

appreciation is greater, but as before, such exchange rate appreciation does not mean any loss of competitiveness, because it proceeds to the point where the domestic price of tradables equals their international price. M=const means

Pˆn  Yˆ

0 , so that the nominal price level should fall by as

much as the total output would grow as a result of productivity increase (or by as much as the relative price of non-tradables increases, as one means the other). The total output growth associated with a 1% increase in pS is

§ M · § 1 ·  >T S  (1  J S )(1  T S )@¸¸  J S ˜ ¨¨  Q ¸¸ \ %(10.34) (1  J S ) ˜ ¨¨ © P NS ¹ © P NS ¹ Then Pn should fall by ȥ%, which means that the numeraire drops by (1-Ȗs+ȥ)%, which is more than in the previous case. Again, the nominal exchange rate should appreciate by the same percentage. There is also a wide range of exchange rate/inflation tradeoffs, in which the money stock grows faster than output, which leads to less exchange rate appreciation but more inflation. Suppose M/Y is not constant, but rather

Kˆ k  (1  J S ) pˆ s .

Since it still holds that

HˆS ,i

PˆTi  PˆTS

Pˆn

k ! 0 . Then PˆTi  Kˆ and we

continue to assume that there is no price changes in country i, İS,i appreciates by ˆ s . This implies that whatever growth money stock may have, exactly k  (1  J S ) p the exchange rate is a mirror image of the movements in the numeraire, therefore, there is no loss of competitive advantage for the countries experiencing the HBS effect, but no gain either. So far, we have seen that productivity growth in S has an impact on income, production and consumption, the price level and exchange rate in S, and now aim to explore what ‘spill-over’ effects these changes have on S’s MU partners. 4. TWO COUNTRIES IN A MU: WHAT INFLUENCE DOES THE HBS EFFECT IN ONE COUNTRY HAVE ON THE OTHER? 4.1. Preliminary observations Now that we have studied the implications of the HBS effect for one country, we are ready to examine its impact on the other countries tied to the FGC by a MU and, therefore, by a single monetary policy. All the implications for one country which were derived in real terms are still valid. However, there are some differences in the effects of monetary policy, and some feedback impact from the FGC on their slower-growing MU partners. In particular: x The fall in the numeraire required to keep the aggregate price level unchanged is smaller now, because the SGC do not experience an increase in their price levels;

NIKOLAI ZOUBANOV

172 x

The SGC may have its exports (and employment) increase through the increase in demand in its faster-growing partner, but taking full advantage of this requires flexible labor market, namely the ability to index the wage to the numeraire (i.e. the exchange rate) in order to stay competitive; x If labor markets in the SGC are rigid, not only will it be difficult for the SGC to export more, but also employment may fall – unless there will be adequate growth in labor productivity; x However, if there is no productivity growth in the SGC and they are unable to make their labor market flexible enough, the union’s Central Bank might consider tolerating higher inflation for the union as a whole (although this would not increase inflation in the SGC). To illustrate this formally, consider a slower-growing country named G forming a MU with S where productivity grows faster than in G. Let S produce one tradable good, AS, which it can export worldwide and sell at the world market’s prevailing price. G produces two tradable goods: AG (identical to AS) and BG, which S cannot produce. Goods AS, AG and BG are also produced and traded worldwide. Both G and S are small countries relative to the size of the world market, so the price elasticity of demand for their exports is assumed to be infinite. All the assumptions made earlier for S are valid for G, except that G does not experience productivity growth at all. Again, this is not a crucial assumption, but does greatly simplify calculations. All the conclusions are valid even if there is some productivity growth in G (lower than in S), although they become less pronounced. In addition to the assumptions made in Section 10.3, we now introduce utility measures for the whole G+S union and for both tradable goods (A and B). We continue to use CES utility functions. The beauty of the CES utility function is that it allows easy aggregation of the goods entering it, and the price index implied by it can contain aggregated prices of its components indexed in a similar way as is the final price index. Furthermore, since we calculate the effects in percentage points, we are free to choose suitable initial values of relative prices, which greatly simplifies calculations and serves for clarity of results. Thus, introducing a CES utility function for the G+S union including aggregate 1

goods S and G,

u (CG  S )

T 1 T

1

T T 1 T 1 T

ª T º T «G ˜ CG  (1  G ) ˜ CS » ¬ ¼

(where į=share of

G’s economy in the union, and ș is the elasticity of substitution in consumption between goods S and G – in fact the propensity to migrate), we derive the G+S aggregate price index in the usual way as

PGS

1 1T 1T

>G  (1G )U @

(10.35)

where ȡ is the relative price of good S in terms of good G. From here we can immediately see that, if a 1% increase in pS causes a (1-Ȗs)% increase in PS, then it also causes a (1- į)(1-Ȗs)% increase in PG+S, by analogy with equation 10.20). Thus

UNEVEN GROWTH IN A MONETARY UNION

173

the presence of a SGC in the MU will attenuate the exchange rate realignment required to keep the nominal price level constant. Another lesson is that the HBS effect in S does not have any bearing on inflation ˆ S , so that the nominal price of tradables in G. As was shown previously, yˆTS w stays unchanged. Thus, there will be no HBS effect-caused price growth in G because of trade with S, no matter how much the price index in S appreciates due to the HBS effect (as long as there is no autonomous HBS effect in G itself). In this section, implications of growth in the FGC for the SGC will be derived only for the “M/Y=const” case. However, one can use previous intuition and insights from Section 10.3 to see that in the case of fixed money stock, the implications to be derived will be qualitatively the same, except that the magnitudes of deflation and exchange rate appreciation will be bigger. 4.2. The impact of growth in S on output and employment in G: the flexible labor market case We assume that in the presence of flexible labor markets the nominal wage can be instantaneously adjusted so as to render the real wage equal to marginal product of labor. However, we do not assume here that there is full employment. Even if the wage is perfectly flexible, there may be structural and frictional unemployment caused by various side factors. Furthermore, there may be a part of population able but not willing to work whose preference towards employment may change over time. So, if there is an increase in the labor demand, it can, in principle, be met without the real wage going up. P WP=MC+T MC effective price of export

D(S+G) domestic consumption export volume

Q

Figure 10.1. The impact on output of BG of an increase in demand for it in S

NIKOLAI ZOUBANOV

174

Recall from previous derivations that a 1% increase in ps results in a

M  (1  J S )(1  T S ) % P NS

increase in demand for tradables, including good BG.

How does G benefit from the increase in demand for BG in S? If G is a net exporter of good BG and faces constant long-term marginal costs MC (because of CRS in the production function), transportation costs for delivery outside the G+S union, T, the world market’s price level, WP, and the aggregate downward-sloping intra-union demand curve, DS+G, then the picture looks as follows (Figure 10.1). G sells BG worldwide at WP (and covers transportation costs), and within the union at MC. The volume of exports to the world is unaffected by the increase in demand within the union, so the union consumes the output of BG that is left. Because of transportation costs, every increase in demand for BG in the union is met by increased production of BG in G, not in the world. Then for every 1% increase in ps G gains

M  (1  J S )(1  T S ) % in employment (if there is a CRS technology P NS

in place) in the sector meeting S’s demand for BG and the corresponding increase in income. 4.3. The impact of growth in S on nominal price levels As we have seen in sub-section 10.4.1, a 1% increase in pS causes a (1- į)(1-Ȗs)% increase in PG+S. Therefore, in order to keep the overall nominal G+S price index constant, the numeraire in G+S (i.e. the price of its tradable goods) has to go down by the same percentage, which implies a corresponding nominal exchange rate appreciation, but no loss of competitiveness for either S or G, just as in the case of one country. However, because of differences in growth rates, we now have positive dynamics in S’s price level and negative dynamics in G’s. Namely, PS rises by (1-Ȗs)(1-į)(1-Ȗs)=į(1-Ȗs)% (because it is cushioned by its slower-growing neighbor), but PG goes down by (1-į)(1-Ȗs)%. 4.4. The impact of growth in S on output and employment in G: the rigid nominal wages case Nominal wage rigidity can be a real threat for a SGC which is in a MU with a FGC, especially when the union’s monetary authorities adopt a strong anti-inflationary stance. A rigid labor market in G means that while PG goes down, the nominal wage does not, thus leading to real wage increases in G unmatched by productivity growth. Therefore the demand for labor will decline. Because of subsequent nominal exchange rate appreciation in the MU currency, there will also be a drop in G’s exports to the rest of the world (as MC in the production of BG in the world’s currency, e.g. USD, shifts upwards) and an increase in its imports, as marginal cost in the production of AG increases and that in the production of AS remains constant.

UNEVEN GROWTH IN A MONETARY UNION

175

All this will increase G’s unemployment and worsen its current account. True, these negative effects will be mitigated by real earnings growth, but the per-capita earnings growth for those in employment must, in its turn, be set against the drop in employment. For G to avoid job cuts and a decline in income from exports, the central bank of the G+S MU would have to consider issuing more money than would be enough to satisfy M/Y=const. A good choice would be to set

Mˆ  Yˆ

(1  G )(1  J S ) for each 1% increase in pS, in which case the price level

in G would stay unchanged, although of course, at the price of somewhat higher inflation in the union as a whole. We next examine the impact of wage rigidity for employment and output in detail in both the short and long run. 4.4.1. Short-run impact We use two definitions of the short run: (i) as the period in which the capital-tolabor ratio is fixed at its historical level, and (ii) as the period in which the stock of capital is fixed. Both definitions lead to fundamentally the same results. Namely, in the short run, without an adequate growth in labor productivity, the increase in the real wage renders G’s exports uncompetitive on the world market, no matter how labor intensive they may be. In this situation, G faces the choice of either withdrawing from the world market or subsidizing the exports of the industries that have costs higher than revenues. However, because of the transportation costs barrier, domestic production will not stop completely. To see this, consider production of goods AG and BG for export outside the union. In the case of fixed K/L ratio, the CRS Cobb-Douglas production function can be transformed as follows:

Y

D

AK L

1D

A kD L

(10.36)

where k=K/L, which is fixed. The total cost function in this case is trivial:

C ( w, r , Y )

L(rk  w)

rk  w Y A kD

(10.37)

from which we derive the marginal (and average) cost function:

MC

rk  w A kD

(10.38)

Log-differentiate (2.4.2) to get:

d log(MC ) (1  D ) wˆ

(10.39)

176

NIKOLAI ZOUBANOV

which implies that, under given technology the only way to prevent average costs from growing (i.e. to break even with exports) is to avoid wage increases in the absence of adequate labor productivity growth. If the wage grows unmatched by productivity, exports become uncompetitive and the workers lose their jobs. If there were some productivity growth, equation 10.39 would look differently:

d log(MC )

wˆ (1  D )  Aˆ

(10.40)

So, it would be possible to offset the increase in costs due to real wage increases by sufficient improvements in labor productivity. Notice here that for a given productivity growth in S’s tradable sector it takes much lower (but non-zero!) productivity growth in G’s tradable sector to restore G’s competitiveness. With a 1% increase in ATS, pS goes up by P TS % (equation 10.16), then the wage in G P NS  1 increases by PTS (1- į)(1-Ȗs)%, which requires PTS (1- į)(1-Ȗs)(1-D)% increase in A P NS P NS in the production function, which is much less than the 1% improvement in productivity in TS. Therefore, G can cope with the impact of productivity growth in S, if it ensures that its workers receive exactly their marginal product. When capital stock is fixed, we come to fundamentally the same results, with the difference that it is now possible to trim marginal costs down to the WP-C level by decreasing the output of BG and substituting labor by relatively cheaper capital. However, with the real wage increase unmatched with productivity growth, it is impossible to restore average costs, so that the export industry will be making negative profits. Consider the output per unit of capital,

y Y / K A(L / K)1D Al1D

(10.41)

where l=L/K. Derive l from equation 10.41 and substitute the result back to the total cost function per unit of capital: 1

c C/ K wl r r  w(y / A)1D

(10.42)

Derive the marginal cost function and log-differentiate it (holding A and r are fixed):

UNEVEN GROWTH IN A MONETARY UNION

177

D

MC

1 w 1D ˜ y 1D 11D A

(10.43)

D ˆ MCˆ w yˆ 1D We see that it is now possible to manipulate output in order to keep marginal costs unchanged. To keep following must hold:

MCˆ

0 in the absence of productivity growth, the

D

ˆ  w yˆ 1D

(10.44)

i.e. with every 1% increase in wage, output must contract by our production function, is equivalent to

1

D

1D

D

%, which, given

% drop in employment6.

Log-differentiating the average cost function, D

r w AC  1 y1D , y 1D A

(10.45)

we get: D

r w 1D ˆ  yˆ  yˆ (1D)w ˆ  y ACˆ (1D)w y 11D A

(10.46)

So, it is not possible to adjust the average costs to their previous level, unless there is an adequate growth in labor productivity. If there is none, then the exporting industries will make losses and, if not subsidized, will have to withdraw from the market. However, if we assume that capital, once allocated to production in a particular sector cannot be reallocated to another (so that it effectively becomes a bye-gone cost), then exports of BG will not collapse immediately. What will collapse is returns to the owners of capital used in the production of BG, who will effectively suffer a capital loss. Such a situation will lead to a slow decline in BG production, as capital used in the sector declines through depreciation which is not replaced.

NIKOLAI ZOUBANOV

178

For intra-union exports of AG the situation is similar: since S is now the most efficient producer of A, the intra-union trade of A will decline, until it ultimately becomes negligible. However, with intra-union sales of BG the situation is different: because of the protection offered to G’s producers by transportation costs of B from the rest of the world, it is cheaper for S to buy BG than its analogue on the world market. This implies that there exists a finite price elasticity of exports on the intraunion market, -ı% per each 1% increase in price. Thus the producers of BG meeting the demand from S do not have to shut down or ask for subsidies. Nevertheless, they will also experience a decrease in output and employment. With each 1% increase in pS marginal costs of good BG increase by (1  J S )(1  G )(1  D c) % (where (1-Ȑ) = the share of labor in the production of BG). The increase in the price of good BG relative to the price of AS (not of AG because there is no point in buying A at a dearer price from G than it can be obtained domestically in S) is also (1  J S )(1  G )(1  D c) %. Assuming a CES demand function for goods A and B in both countries, the demand for B from S falls by Z  (1  E )(1  Z ) (1  J S )(1  G )(1  D c)% , where Ȧ is the elasticity of

>

@

substitution in consumption between A and BG and ȕ is the share of BG in total consumption of tradables by S. However, with income in S rising, the net increase in demand for BG is

O

M  (1  J S )(1  T S )  >Z  (1  E )(1  Z )@(1  J S )(1  G )(1  D c)% P LNS (10.47)

per 1% increase in ps, the sign of which is ambiguous. One insight of the expression above is that if good BG is capital intensive then nominal wage rigidity in G does not have too great an impact on demand for BG in S, but that otherwise G is likely to lose sales even in “friendly” S. Recalling equation 10.41 for BG, we obtain the net increase in employment in sector BG, Ȝ/(1-Ȑ)%, while in sector AG there is clearly negative employment growth. So, the only hope for G’s exports to S to recover is for higher exports of BG, which is not produced in S and which it is cheaper for S’s consumers to buy from G than to import from the world market (up to the point at which the increase in the price of GB corresponds to the transportation cost of bringing B from the world market). With employment and output in export sectors decreasing in G, the trend in domestic sales is also likely to be downward (the detailed calculations are skipped here). Increases in wage costs will lead to higher unemployment, which will further reduce consumption. True, the increase in per-capita labor income may stimulate demand and thus reduce the layoffs, but it will not completely offset the divestment of labor, because of the substitution effect between labor and capital. Therefore, although the situation with output is unclear, in the end we observe a decrease in employment in the sum of the tradable and non-tradable sectors, except as regards BG, where the situation with employment is uncertain.

UNEVEN GROWTH IN A MONETARY UNION

179

4.4.2. Long-run impact In the long run, capital use can be adjusted to its most efficient level. With wages remaining rigid, and in the absence of labor productivity growth, G will lose its competitiveness on the world market, because there will be no way to adjust its marginal costs back to the world market’s price. Changing the technology parameter, ‘alpha’, will not reduce costs either. Since it is hardly possible that the government will subsidize G’s industry in the long run, inability to adjust marginal costs down will result in withdrawal of G’s goods from the world market. Of course, this is in case of infinite price elasticity of exports. In another framework (e.g., Dixit and Stiglitz, 1977), where countries are assumed to have some pricing power on the world market, higher costs would mean a reduction of exports, not their complete cessation. On the intra-union market, the situation will be the same as described for the short run. To see why marginal costs cannot be adjusted downwards with real wage increases and zero productivity growth, consider the cost minimization problem with a Cobb-Douglas production function with CRS:

min wL  rK , s.t. AK D L1D

Y

(10.48)

After some algebraic manipulation, we obtain the demand functions for capital and labor:

K

Y ª D wº ˜ ˜ A «¬1  D r »¼

1D

,

Y ª1  D r º ˜ A «¬ D w »¼

L

D

(10.49)

and then the total cost function: D

C

Y ªrº ª w º ˜ A «¬D »¼ «¬1  D »¼

1D

(10.50)

and the marginal cost function:

c

1 A

D

ªrº ª w º «¬D »¼ ˜ «¬1  D »¼

1D

(10.51)

Marginal costs, in this case equal to average costs, are now constant regardless of the volume of output, so that we cannot simply reduce them by reducing output. Changing the “alpha” does not solve the problem, either. To see this, logdifferentiate equation 10.51 holding A and r fixed to get:

NIKOLAI ZOUBANOV

180

cˆ

D º ª Dˆ DDˆ  (1  D ) ˜ « wˆ  1  D »¼ ¬

(1  D ) ˜ wˆ

(10.52)

Again, as in the short-run case, if there were adequate labor productivity growth, the marginal costs would not increase, so that G would remain competitive in the world market. This can be seen immediately from equation 10.51. 4.4.3. Other implications. The wage rigidity in G has further repercussions for its economy, especially for income and employment. Acquiring extra capital will worsen G’s current account. Domestic demand in G will be damaged by unemployment: although there will be an increase in real wages, some labor will be divested thus lowering aggregate demand. Some second-order effects (e.g., demand switching from AG to cheaper AS, labor migration from G to S, market segmentation) could also be modeled. But it does not take an advanced modeling exercise to make the important point here: in SGC, wage rigidities seriously damage the economy by reducing employment and income, and depriving citizens of a great deal of the potential benefits that fast growth in FGC could bring. Therefore governments in SGC should not focus on trying to delay the accession of new FGC or on hindering growth accelerating reforms in their current partners, but rather on institutional reforms to enhance the flexibility of their own markets. An alternative solution to the dilemma of the SGC with inflexible labor markets, is for the MU central bank to allow slightly higher inflation in the union. However, this would not be for the sake of faster-growing S (it is doing fine at whatever the union-wide inflation rate), but to help G’s workers keep their jobs, in spite of their nominal wage rigidity and the ensuing mismatch between real wages and productivity in G. Such an increase in the MU inflation, however, would not have to be met by increases in the interest rate, as long as price increases in non-tradables in S are caused by productivity (and therefore production) growth in S’s tradables sector. 5. CONCLUDING REMARKS The results obtained above suggest the following: 1. For a MU, having countries that grow at a faster-than-average rate does not necessarily increase union-wide inflation, if a neutral monetary policy stance (M/Y=const) is adopted by the central bank. However, this comes at the expense of higher inflation in FGC and lower inflation (or even deflation) in SGC of the MU. 2. Under the neutral monetary policy described above, and provided that SGC have flexible enough labor markets (so that they can adjust their nominal wages in line with actual inflation or deflation), SGC clearly benefit from increased demand for their exports from FGC. If, however, there are rigid labor markets in SGC, they are likely to suffer a fall in output and

UNEVEN GROWTH IN A MONETARY UNION

181

employment. Even though there will still be an increase in demand from FGC, it’s effects on output and employment in SGC may be significantly limited by increased real production costs. For the same reason SGC, with rigid labor markets and fast growing co-members in a MU implementing a neutral policy stance, will find it increasingly hard to compete on the international markets with more efficient producers – unless they generate adequate productivity growth. However, the amount of productivity growth that needs to be generated for SGC to avoid suffering as a result of fast growth among their MU partners is a fairly small fraction of the growth that the FGC achieve. 3. Therefore, the real problem SGC face in admitting new FGC to a MU they belong to, or seeing current members growing faster, is not that this will increase union-wide inflation (let alone the SGC domestic inflation), given a neutral monetary policy, but rather that their rigid labor markets will render them less competitive both union-wide and internationally. Generally, the more labor-intensive their exports are and the more oriented they are to world trade outside the MU, the more will their losses in employment and output be. 4. The overall gain or loss for SGC from the MU’s expansion depends on many factors, such as technology parameters (labor intensity), shares of tradables and non-tradables in consumption, the share of exports going to countries within the MU, nominal rigidities in labor markets, etc. However, the MU central bank may want to protect SGC from losses by setting its monetary policy so that their aggregate price level is unchanged (most importantly, does not decline). In this case there will be no nominal appreciation of the MU currency, no losses due to labor market rigidities, but higher union-wide inflation. However, it is important to understand that this measure is undertaken to protect SGC, while FGC would do just as well under a zero inflation target. There are also a few policy implications following from the theoretical discussion above. First, in a MU in which members experience unequal productivity growth rates and have rigid labor markets, the monetary authorities are faced with three choices (and their combinations, wherever possible): (i) low or zero unionwide inflation, with output contraction and unemployment in SGC; (ii) higher inflation in FGC, low or zero inflation in SGC, and moderate inflation union-wide, with no losses of income or employment in SGC; (iii) zero union-wide inflation and labor market reforms in SGC. The final choice depends on the decision-making process and on how successfully individual countries lobby their positions with the monetary authority, which is supranational by default. Thus, the first choice seems to be the best for the MU Central Bank, as it pursues union-wide objectives, such as the aggregate inflation level. The third choice is also good for the CB, but individual countries might find it difficult to implement labor market reforms. The second choice is the best for the SGC, which do not have to undertake any reform in this case, but the CB might well argue against it and FGC might loose slightly from the slightly higher inflation they experience.

182

NIKOLAI ZOUBANOV

In the end we would like to stress that in no case should higher-than-average productivity growth be an obstacle for any country trying to join a MU. High growth in some MU members is an economic blessing that can be shared among all the union’s participants, if they are flexible enough to benefit from it, and not a curse that should be avoided. With long-established economic ties and political integration accelerating recently, fast and smooth accession of the NMS to the EMU may become critical for the overall success of the European project.

NOTES 1

I would like to thank Jacek Rostowski for his extensive comments and valuable advice in researching this topic. 2 The same sort of calculation can be performed for each 1% increase in A using equation 10.16. TS 3 However, the growth of total output is much less important to us than the growth of tradable output, because it is tradable output that influences the current account, which in its turn, may induce exchange rate realignment to keep the current account constant. So, the exchange rate may be influenced by productivity growth not only directly, but also indirectly, through the current account balance. Exploring this link is relevant for our topic, but is left out of this chapter for further exploration. ˆ 4 Recall that C and Cˆ NS . So, for a positive growth M TS M  >T S  (1  J S )(1  T S )@  (1  J S )(1  T S ) pˆ s pˆ s P NS P NS in consumption of tradables, M t (1  J )(1  T ) ˜ P ; and for a positive growth in consumption of nonS

S

NS

tradables, M t >T  (1  J )(1  T )@ ˜ P . S S S NS 5 In a more complex setup, though, when we have to deal with non-tradable inputs in tradable goods, there may be some loss of competitiveness, but the question then is whether such goods should be classified as fully tradable. 6 Consider that lˆ Lˆ  Kˆ Lˆ .

REFERENCES De Grauwe (1992). Inflation convergence during the transition to EMU. Centre for Economic Policy Research (CEPR) discussion paper series, 658. Dixit, A. & Stiglitz, J. (1977). Monopolistic competition and optimum product diversity. American Economic Review, 67, 297-308. Gordon, Robert J. (1998). Is there a tradeoff between unemployment and productivity growth? NBER Working Paper, 5081. Obstfeld, M. & Rogoff, K. (1996). Foundations of international macroeconomics. The MIT Press, chapters 2 & 4.

WOJCIECH PACZYNSKI

CHAPTER 11

ECB DECISION-MAKING IN AN ENLARGED EMU

1. INTRODUCTION The 2004 enlargement and the prospect of the new member states (NMS) adopting the common currency in future years have motivated the EU to undertake several important changes in its institutional design. The primary difficulty in adopting such reforms results from the conflicting objectives of promoting greater efficiency in decision-making at the EU level and securing the rights of sovereign member states (MS) to make their own decisions on matters they consider important. The dilemmas faced on the occasion of the recent enlargement are arguably greater than was the case in the past due to the very large number of new entrants, their different characteristics in comparison to the old member states (OMS) and the much higher level of political and economic integration that has been achieved in the EU during the 1990s. In particular, the enlargement of the EMU is often seen as a major challenge for the European Central Bank (ECB) which implements monetary policy in the Eurozone. This is the starting point of the analysis in this chapter. The main claim is that the arguments calling for ECB reform tend to overestimate the potential negative impact of the enlargement. Among the proposed reform options there is little firm evidence that clearly supports one option over any other, but several solutions appear reasonable. The reform recently adopted by the EU Council, while not free from shortcomings, should nevertheless provide a reasonable environment for efficient monetary policy-making in the enlarged EMU. 2. DECIDING ON POST-ENLARGEMENT EMU MONETARY POLICY This section summarizes the discussion on the need for reform of decision-making processes within the ECB after the Eurozone enlarges from its current 12 to 20, 25 or more MS. As stipulated by Article 107 of the Treaty, the European System of the Central Banks (ESCB) is governed by the decision-making bodies of the ECB, i.e. the Governing Council (GC) and the Executive Board (EB). The EB comprises six members (President, Vice-President and four other members). As stated in Article 112 of the Treaty, all Board members

183 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 183-198. © 2006 Springer. Printed in the Netherlands.

184

WOJCIECH PACZYNSKI ‘shall be appointed from among persons of recognized standing and professional experience in monetary or banking matters by common accord of the governments of the Member States at the level of Heads of State or Government’1.

The GC is composed of members of the EB and governors of the national central banks (NCB) of EMU MS. All members of the GC have one vote and in most cases decisions are taken using the simple majority principle. In the event of a tie, the President has the casting vote. In particular, interest rates decisions are taken using this principle. As in the case of other EU institutions, enlargement of the Union was generally viewed as necessitating a reshaping of decision-making rules in the ECB. However, no consensus was reached during the Intergovernmental Conference (IGC) leading to the Nice Treaty, which only called for the ECB or the Commission to propose reform ‘as soon as possible’. Two major arguments for reform are put forward in the context of enlargement. First, it is argued that the expected Eurozone enlargement will significantly increase the share of small, fast-growing economies in the EMU. It is further feared that the preferences of many GC members might be affected by conditions in their home countries. If this were the case, the (distorted) views of representatives of smaller, catching-up economies might prevail in the GC and thus monetary policy might fail to reflect the needs of bigger EMU economies (the ‘core’) and thus of the EMU as a whole. Second, it is claimed that the sheer size of the body deciding on monetary policy (GC) will become too large to ensure efficient and meaningful discussion of economic developments in the EMU, and that as a result the ability to manage the Union’s monetary policy might be affected. Indeed, by around 2010 the Eurozone might well comprise 25 or so MS, which would bring the number of people in the GC to around 30. 3. EVALUATION OF THE SMALL COUNTRIES’ BIAS ARGUMENT After enlargement the number of small economies in the Eurozone will increase significantly. Macroeconomic diversity within the EMU will also widen, with most of the acceding small countries expected to exhibit higher average growth rates than the EMU core economies, and also higher average inflation rates. There is a distinct question on the expected sources of the inflation differentials and how they should be tackled by the ECB. Such issues are discussed in more detail by, for example, Honohan & Lane (2003), ECB (2003), Angeloni & Ehrmann (2004) and in Chapters 1 and 10 of this volume. Here we only observe that such differences might, under some circumstances, turn out to be important in the decision-making process, providing that GC members put more weight on the needs of the countries they come from than would be justified by their respective country’s economic weight in the union. Such a phenomenon is referred to as the ‘regional bias hypothesis’. Article 108 of the Treaty demands that GC members act independently in fulfilling their duties. Baldwin, Berglof, Giavazzi & Widgren (2001) comment that ‘A Panglossian observer would be satisfied with this; a Machiavellian observer would laugh. The truth is probably somewhere in between’2.

ECB DECISION-MAKING IN AN ENLARGED EMU

185

Since the discussion and voting results within the GC are kept secret, there is no direct way to test the regional bias hypothesis. A narrative approach, looking at whose interests were best secured during the initial period of the EMU’s functioning, while subject to strong methodological critique, does not provide decisive results (Baldwin, Berglof, Giavazzi & Widgren, 2001). In this context the US Fed represents an interesting object of study of regional representatives’ voting behavior. Most of the studies (see Meade & Sheets, 2004 for a survey) found mixed and not very robust results. In a recent work Meade & Sheets (2004) illustrate a relationship between regional unemployment rates and attitudes to interest rate decisions by the Federal Open Market Committee (FOMC) members from particular regions. Evidence consistent with regional influences is found among both regional Fed Presidents and Board members and is actually stronger among this latter group. It is interesting to note that this finding raises doubts about the popular view that Frankfurt-based EB members will tend to care about EMU-wide developments while it is CB governors who will tend to take a national perspective. Such an assumption underpins all the scenarios for hypothetical coalition formation in the GC presented by Baldwin, Berglof, Giavazzi & Widgren (2001) or Eichengreen & Ghironi (2001). There is little doubt that with the NMS adopting the Euro, the economic heterogeneity of the EMU will increase significantly. Nevertheless, the conclusion that this poses risks to GC voting outcomes, should NCB governors exhibit regional biases, is not straightforward. Below we analyze the historical inflation patterns among current and future members of the Eurozone and assume that this provides a realistic prediction of inflation rate dispersion in the future. The main finding emerging from this exercise is that while the enlargement of the Eurozone is indeed likely to widen the inflation dispersion among the MS (compared to the smaller Eurozone of today), median inflation is unlikely to change much. Moreover, once we assume six EB members base their decisions solely on EMU-wide inflationary developments, the outcome of majority voting appears to mimic very well the outcome of decisions based on the EMU-average inflation only, even if we allow NCB governors to exhibit regional biases in their most extreme form. Clearly, past inflation experience might not provide good predictions for future developments after the NMS join the EMU, but the size of inflation differentials among countries analyzed in this exercise (in excess of 10 percentage points during most of the 19982000 period) appears very high in comparison to inflation dispersion expected after Eurozone enlargement. One simple argument supporting theview that dispersion will not be higher than during 1998-2000 is that if, after adoption of the Euro, inflation in the NMS continued at its 1998-2002 levels, price levels in these countries would very quickly (within 5-10 years) go well above the EU-15 average (see Maier & Cavelaars, 2003).

186

WOJCIECH PACZYNSKI

4. MODELING ECB GOVERNING COUNCIL DECISION-MAKING IN AN EMU OF TWENTY-FIVE Let us concentrate on the worst-case scenario, which gives rise to the most alarmist calls for reform. We consider a one-stage game in which the distribution of national inflation rates is given and the question is what interest rates will be set as a result of the majority voting rule. We make the following simplifying, though arguably not unrealistic, assumptions: 1. the only parameter that ECB EB members take into account to form their private preferences about monetary policy is the Eurozone inflation rate; 2. the only parameter that NCB governors take into account to form their private preferences about monetary policy is the inflation rate in their home economies; 3. the only monetary policy tool that the ECB collectively decides on is the interest rate; 4. there is a monotonic positive relationship between observed inflation rates and preferred interest rates; 5. each member of the GC has its unique preferred interest rate level (bliss point) R*, and for any choice of two alternative interest rate levels either below or above R* he prefers the one closer to R*. Assumption (e) is the so-called single-peakedness assumption (Black, 1958; for an exposition see, for example, Myerson, 1996). It ensures that a version of the median voter theorem holds. Specifically, under such conditions, if the number of voters is odd, then the median voter’s bliss point is a Condorcet winner. This means that under normal conditions (i.e. excluding limitations to the voting agenda and manipulation of the agenda), the GC would adopt the interest rate which is equal to the median of interest rates preferred by the players involved. In order to get more insights from this very simple result one needs to have some idea on the distribution of interest rate preferences among members of the GC. In the model discussed this boils down to the question of the distribution of inflation rates in the EMU. Given assumptions (a)-(d), the median preferred interest rate will be determined by the median inflation rate (the median is calculated for the set of inflation rates used to form interest rate preferences by GC members). The first interesting insight is provided by comparing median inflation rates in the EMU-12, EMU-25 and the median of the four largest economies of the hypothetical EMU-25 (Figure 11.1). The EMU-25 median remained slightly above the EMU-12 median throughout most of the period, with the difference rarely exceeding 0.3 percentage points (apart from 1998) and the opposite relation prevailed during the last two years for which data were available. Also, both the EMU-12 and EMU-25 medians remained above the Big 4 median throughout most of the period.

ECB DECISION-MAKING IN AN ENLARGED EMU

187

4 3,5 3 2,5 2 1,5

EMU25 EMU12 Big4

1 0,5

19 98 1 9 m01 98 19 m0 5 98 1 9 m09 99 19 m0 1 99 1 9 m05 99 20 m09 00 2 0 m01 00 20 m05 00 2 0 m0 9 01 20 m01 01 20 m0 5 01 2 0 m09 02 20 m01 02 2 0 m05 02 20 m09 03 2 0 m0 1 03 20 m05 03 2 0 m0 9 04 20 m01 04 2 0 m0 5 04 m 09

0

Figure 11.1. Median inflation rates (HICP) in EMU-12, (hypothetical) EMU-25 and in the four largest EMU-25 economies, January 1998- September 2004 (annual change, % points) Notes: The “Big 4” group includes France, Germany, Italy and UK. Source: Own calculations based on Eurostat data.

3,5 3 2,5 2 1,5 1 0,5

EMU media n( 12+ 6) EMU 12

19 98 m 19 01 98 m 19 05 98 m 19 09 99 m 19 01 99 m 19 05 99 m 20 09 00 m 20 01 00 m 20 05 00 m 20 09 01 m 20 01 01 m 20 05 01 m 20 09 02 m 20 01 02 m 20 05 02 m 20 09 03 m 20 01 03 m 2 0 05 03 m 20 09 04 m 20 01 04 m 20 05 04 m 09

0

Figure 11.2. Median inflation in the EMU-12 compared to EMU12 inflation, January 1998September 2004 (annual change, % points) Notes: Median for EMU (12+6) is defined as a median in the set which includes six times the HICP of the EMU-12 (to reflect the assumed preferences of EB members) and national inflation rates in each of the 12 EMU MS. Source: Own calculations based on Eurostat data.

WOJCIECH PACZYNSKI

188

We now turn to analyzing GC voting outcomes regarding interest rates under the assumptions formulated above. These will be determined by median inflation rates, which we calculate based on a set of 18 ‘voters’ in the case of the current EMU-12, consisting of the 12 EMU countries and six voters (the EB) who vote on the basis of ‘Eurozone’ inflation. In the case of our hypothetical EMU-25, we have 31 voters consisting of the EMU-25 member states and the 6 EB members voting on the basis of EMU-25-wide inflation. Figures 11.2 and 11.3 plot these median inflation rates for the period 1998-2004, comparing them to union-wide inflation rates. A striking observation is that throughout the period analyzed, apart from short episodes, median inflation calculated in this way coincided with aggregate inflation, both in the case of the EMU-12 and of the EMU-25. Under the assumptions formulated above, this would assure that monetary policy decided by the GC would serve the interest of the EMU as a whole quite well, whether it is composed of 12 or 25 MS. 3,5 3 2,5 2 1,5 1 0,5

EMUmedian(25+6) EMU25

19 98 19 m01 98 19 m0 98 5 19 m09 99 1 9 m0 99 1 1 9 m05 99 20 m09 00 2 0 m0 00 1 20 m05 00 2 0 m0 01 9 2 0 m0 1 01 20 m05 01 2 0 m0 02 9 20 m01 02 2 0 m0 02 5 20 m0 9 03 20 m01 03 2 0 m0 03 5 2 0 m09 04 20 m01 04 2 0 m0 04 5 m 09

0

Figure 11.3. Median inflation in a hypothetical EMU -25 compared to EMU-25 inflation, January 1998- September 2004 (annual change, % points) Notes: The median for the EMU-25 is defined as a median in the set which includes six times the HICP value for the EMU-25 and national inflation rates in each EMU-25 MS. Source: Own calculations based on Eurostat data.

Clearly, one should be aware of the limitations of this exercise. In reality, interest rate preferences are based on a large set of indicators such as output gap, forecasts and projections of future inflation paths, etc. rather than current inflation levels only. Also, the assumption of a monotonic relationship between inflation levels and preferred interest rates is a simplification – 3% inflation in Germany might call for more restrictive monetary policy than 4% inflation in Slovakia (should both

ECB DECISION-MAKING IN AN ENLARGED EMU

189

countries vote for interest rates optimal from their own perspective). However, this last argument actually suggests that the dispersion of interest rate preferences should be lower than the dispersion of inflation levels. Furthermore, it is worth bearing in mind that the sample analyzed includes a period of very high (double digit) inflation in several NMS (1998-2000) resulting in the dispersion of inflation rates among the EMU-25 group members that is unlikely to be repeated after adoption of the common currency – as can be judged from the record of the period during 20012004 and from the experience of those NMS that maintained fixed exchange rates. All in all, these results suggest that even in a large and very heterogeneous MU just a few members of the interest rate setting body adopting the union-wide perspective should make a system similar to the one currently in operation at the ECB resilient to even very strong potential regional biases. The ECB GC decides on the interest rate rather than on inflation. Because the link from one to the other is complicated, dependent on other macroeconomic variables, subject to uncertainty and varies between countries and over time, it is likely that the dispersion of national preferred interest rates is smaller than that of observed national inflation rates. One argument is that a (perhaps substantial) part of the difference in inflation rates might reflect natural adjustment processes (for example resulting from varying productivity growth rates) and therefore should not be counteracted by economic policies3. Another observation is that due to strong economic linkages between the countries of the current and future enlarged EMU, national preferred interest rates should incorporate the preferences of other partners in the MU. If an important trade partner badly needs lower interest rates, this should impact on domestic objectives – even if only for purely self-interested reasons. This mechanism should also act towards decreasing the dispersion of preferred interest rates4. In conclusion, we have illustrated that observed inflation differentials do not lead to a substantial divergence of the median inflation rate from the EMU average both in the EMU-12 and in the EMU-25. Second, we have illustrated that the small countries bias should not pose a threat to the conduct of ECB monetary policy even if regional preferences are present. The six EB members basing their assessment on EMU-aggregate inflation should prevent the hypothetical ‘fast growing, high inflation’ group of countries from gaining a majority in the GC after EMU enlargement even unchanged decision-making rules. However, it should also be mentioned that whatever the truth about the functioning of the GC is, its perception by financial markets also matters. If the GC is perceived as inefficient in managing EMU monetary policy and subject to regional biases, this would be bad for the Union irrespective of the actual decisions of the GC. By the same token, even if internal discussions were difficult and some members overemphasized local conditions, it would not necessarily have a negative impact on the functioning of the Union as long as the credibility of the ECB was maintained.

190

WOJCIECH PACZYNSKI

5. AN EVALUATION OF THE GOVERNING COUNCIL SIZE ARGUMENT The ECB Governing Council is already rather large. There are 18 members voting on interest rate decisions compared to 12 voting members of the US FOMC, and between 6 and 10 in the monetary policy committees (MPC) of Sweden, Canada, Australia, the UK and Poland. However, prior to German unification the Bundesbank Council also consisted of 18 members (7 Board members and 11 presidents of regional banks)5. A further increase in the size of the ECB GC (to around 30) would clearly make it very difficult to conduct open discussions on economic developments in the EMU involving all members of the body6. One argument raised in this context is that there seems to be a culture of long discussions leading to ‘consensus’ decision-making without voting in the ECB GC7. Clearly, even if decisions are taken without a formal vote, this does not mean that all members agree on a particular decision. Once the positions of all members are known, the outcome of a potential vote is established anyway. While the ability of the GC to carry on in-depth discussion would obviously be affected by the enlargement, there are no problems with the decisionmaking procedure itself. Simple majority voting with the threshold set at 50% of all votes assures that there will always be a majority supporting some monetary policy decision8. The problem of the optimal size of the MPC has also been studied from the more formal perspective applying the tools of game theory. This literature, while providing some interesting results under specific sets of assumptions, does not come up with any clear-cut policy recommendations. The optimal size of the committee can be affected, among other things, by factors such as the cost of accessing private information and the heterogeneity of objective functions among committee members. An up-to-date survey of the relevant literature is provided by Gerling, Gruner, Kiel & Schulte (2003). In summary, the issue of the size of the GC does indeed constitute a potential problem for this body’s ability to make swift policy decisions. There is no experience of other CBs with such large decision-making bodies. On the other hand, the ECB itself is a unique institution in global economic history. Proposals to limit the size of the GC would be welcome, provided they fulfill the other criteria discussed below. Whether the GC of 30 or so members could organize its work efficiently remains an open question. 6. POSSIBLE SOLUTIONS Several solutions have been proposed to solve one or both of the above problems. These mainly reflect the experience of existing arrangements in other institutions facing similar challenges. The major ideas implemented (in various combinations) are rotation, representation (forming constituencies) and executive decisions (delegation to technocrats) as shown in Table 11.1.

ECB DECISION-MAKING IN AN ENLARGED EMU

191

Table 11.1. The Pros and Cons of various ECB reform options Criterion / Solution Solving the ‘numbers problem’ Avoiding the ‘small countries’ bias’ Building a common European identity Democratic legitimacy

Rotation of MS Yes a Yes c No Yes c

Representation (constituencies) Yes a Yes c No b Yes c

Weighted voting No Yes No Yes

Small MPC Yes Yes Yes d No

Notes:

a – provided non-voting members do not participate in discussions. b – unless one regards building identity within smaller groups as a step to European identity building (see Heisenberg, 2003 c – debatable, might depend on details of operation. d – unless the MPC is excessively dominated by large countries. Source: own elaboration. Compare Heisenberg (2003).

First, one can think of systems where the composition of the GC remains unchanged, but the rights to vote on policy decisions are given to a subgroup of GC members at any one time. Voting rights would rotate among GC members according to an agreed schedule. This requires that the number of voting members and the period for which they retain voting rights be decided. The advantage of such an approach is that it can limit problems relating to the number of people at the discussion table. However, this is only possible when non-voting members are excluded from discussions, which in turn could have some negative consequences. For instance, members just starting their voting period would not personally intimately know the state of discussion over the previous months or years (even if they had access to minutes from previous meetings). Another solution would be to form groups of central bankers with each of the group delegating a representative for voting (and possibly also discussion). Such a system is used in the IMF Board of Directors9. Under such an arrangement the number of people at the table could be reduced (again, providing that non-voting GC members do not participate in discussions). On the other hand, it would seem natural to treat representatives of groups as representing the interests of their groups rather than being independent experts on EMU monetary policy. It could therefore actually worsen the regional bias problem. One could combine any of the above systems with weighting of countries’ votes by, say, shares in the ECB capital10. This would make the ECB decision-making rules similar to the ones in operation in some other EU institutions, in particular, to qualified majority voting (QMV) in the EU Council. Such a solution would have the advantage of making the regional bias hypothesis irrelevant, since even in the case of purely nationalistic behavior exhibited by GC members, the voting results would still be close to the EMU-wide optimum11. On the negative side, it might strengthen the perception of monetary policy decisions being the outcome of a game between national representatives rather than one motivated by the aggregate needs of the Eurozone. It is interesting to note that a weighted voting scheme for the ECB GC was one of the possibilities discussed during the 1990 IGC on monetary union. The proposal from the European Commission advocating such a solution was at that time

192

WOJCIECH PACZYNSKI

opposed by the Bundesbank and the German Ministry of Finance, who feared that it would ‘encourag[e] a damaging spirit of compromise amongst national interests’ (Heisenberg 2003).

Yet another possibility is to delegate the running of monetary policy to a small MPC consisting of experts, without considering their nationalities. Such a move would be radical as it would constitute a departure from the original ESCB architecture put in the Treaty. It is, however, an actually applied solution in a number of countries such as the UK, which is commonly perceived as an example of extremely successful monetary policy-making (see Begg et al., 2003). This solution would solve both the ‘numbers’ and regional bias problems, provided that members of this body could indeed be regarded as fully independent and not influenced by developments in their home countries. The major disadvantage is that it could undermine the accountability of the ECB. The point is that NCB governors can be perceived as fulfilling the role of a national ‘listening post’, ensuring the ECB is accountable to someone with strong credibility in the home country (Baldwin, Berglof, Giavazzi & Widgren, 2001). Also, there are some potential gains from having national representatives in the decision-making body that would be lost by delegating policy to a small council of experts. CB governors may have better information concerning their economies and better understanding of policy transmission channels. Indeed, in the US regional board governors sitting in the FOMC are regarded as experts on local developments. In the words of Alan Greenspan, ‘As keen observers of local economies, the directors here and elsewhere contribute vitally to the formulation of monetary policy by offering important insights absent, by definition, from even the most careful analysis of aggregate data. Often they know what is happening in the various regions of the country well before the hard data are collected by national statistical agencies’12.

Developments at the level of individual countries are clearly not irrelevant, despite the ECB focus on the Eurozone’s aggregate performance. De Grauwe and Senegas (2003) find that uncertainty about the transmission process increases the need to take into account information about national economies (and not only aggregate data) in the formulation of optimal monetary policies in the EMU. Earlier studies cited there found that asymmetries in the transmission mechanisms of monetary policy across nations also call for a consideration of national data even in the absence of uncertainty. Finally, one might be skeptical about the possibility of identifying and selecting, through a political process, experts able to completely forget their national attachments. The Meade and Sheets (2004) findings on regional biases among FOMC members could be interpreted as lending support to such skepticism. One could note that in a small MPC a large proportion of members are likely to come from the large countries. This is for two reasons: (1) what one might call the ‘Trichet effect’, i.e. simple bullying by the governments of the large countries to make sure that their national representative gets selected; (2) a ‘demographic effect’, resulting from the fact that, with talent being distributed democratically, council members are

ECB DECISION-MAKING IN AN ENLARGED EMU

193

likely to come from countries roughly in proportion to their population (or GDP if that is taken to reflect the ‘human capital’ available to train top economists). This can lead to an MPC excessively dominated by large countries, even in the presence of appointment purely by merit. 7. AN EVALUATION OF THE ECB’S REFORM PROPOSAL In line with the Nice Summit, in December 2002, two years after the signing of the Treaty, but soon after its acceptance by the second Irish referendum, the ECB revealed its proposal for reforming the voting modalities of the GC. On February 3, 2003, just after the Nice Treaty had come into force, the ECB published its official recommendation to the European Council, which accepted the proposal on March 21, 200313. The solution adopted was to introduce a rotation system in a manner that takes into account the economic weight of the MS.

% of meetings with voting rights

80 70 60 50 40 30 20 10 0

'Big 5'

19

2nd group

20

21

3rd group

22

23

24

25

26

27

No. of EMU MS

Figure 11.4. The size of the EMU and the distribution of voting rights Note: The distribution of voting rights is presented only for an EMU of more than 18 MS.

It is foreseen that while all NCB governors should continue to take part in GC discussions, some of them will temporarily not have a vote on interest rate decisions. The six members of the EB will retain their rights to vote at all times, but the governors of national CBs will be allocated no more than 15 votes. The voting rights will rotate within two groups (for an EMU of 16-21 countries) or three groups (if the EMU comprises of more than 21 member states). All countries will be ranked according to their share in EMU GDP (weight 5/6) and their share in the total aggregated balance sheet of monetary financial institutions (‘TABS-MFIs’, weight 1/6). The five biggest economies according to this ranking will form the first group with four votes. In the scenario with two groups, all other countries will be allocated 11 votes between them. When 22 or more countries participate in the EMU, the allocation of voting rights to the Big 5 group will be unchanged (i.e. 4 votes). The

194

WOJCIECH PACZYNSKI

second group will comprise half of all the EMU member states (coming after the Big 5 in the ranking) and will share 8 votes. The remaining countries will have 3 votes at their disposal (see Figure 11.4). Such a procedure clearly introduces some ‘breaking points’ resulting from changes in the ranking due to new economic data and/or inclusion of new countries in the EMU. In many constellations the automatic application of the rule would produce outcomes that were unwarrantedly beneficial or detrimental to some countries or groups of countries. One such scenario, where at some stage there were 16-18 MS forming two groups, with the five biggest economies sharing four votes and the remaining 11-13 countries sharing 11 votes is explicitly excluded by the additional requirement that ‘The frequency of voting rights of the governors allocated to the first group shall not be lower than the frequency of voting rights of those of the second group.’

Also, a kind of a transitory escape clause arrangement is included, stipulating that the introduction of the rotation system might in fact be postponed until the EMU is enlarged to comprise at least 19 countries. It is likely that the rotation system will not be implemented until there are 19 EMU MS, i.e. until seven new countries (out of the 13 current MS not belonging to the Eurozone) join. This appears unlikely before 2008-2009. More generally, some degree of ad hoc decision-making regarding the voting modalities cannot be avoided. These issues will be decided by all GC members – irrespective of whether or not they hold a voting right at the time of the decision – by a two-thirds majority. It is not difficult to see that the proposed system is a compromise, trying to address the contradictory principles of ‘one member one vote’ versus ‘representativeness’ with an attempt to control for ‘automaticity and robustness’. It seems fair to say that, for most sizes of the EMU, it reduces the ‘small countries bias’ present in the current system (where the governor from Luxembourg has the same impact on ECB interest rates as the Bundesbank governor). This perhaps does not go far enough to satisfy everyone in the biggest EMU economies, but has already sparked voices of protests in some small countries. The proposal does not address the issue of the large number of persons participating in monetary policy discussions and decisions. It is hard to believe that there is any difference for the efficiency of the process between: (a) having some 30 members participating in the discussion followed by a vote of 21 and, (b) allowing all members to vote. The hope is that the simple majority decision rule will assure reasonable results in any case14. The complexity of the proposed solution is in itself a reason for criticism. Indeed, the proposed algorithm cannot be described as simple and will demand some degree of ad hoc modification. This might be bad for the transparency of the ECB and public perception of its functioning. On the other hand, it should not come as a real problem for financial markets which have dealt with much more complicated issues.

ECB DECISION-MAKING IN AN ENLARGED EMU

195

Table 11.2. Estimated allocation of MS to groups (according to the methodology adopted by the EU Council and the ranking based on shares in ECB capital)

Group

Group 1 (4 voting rights)

EMU-22 (EMU-12 plus 10 NMS) ranked by criteria adopted by the EU Council Germany France Italy Spain Netherlands

Group 2 (11 voting rights)

Belgium Austria Ireland Poland Portugal Greece Luxemburg Finland Czech Rep. Hungary Slovakia

Group 3 (3 voting rights)

Slovenia Lithuania Cyprus Latvia Estonia Malta

EMU-25 (all current EU MS) ranked by criteria adopted by the EU Council Germany UK France Italy Spain Netherlands Belgium Sweden Austria Denmark Ireland Poland Portugal Greece Luxemburg Finland Czech Rep. Hungary Slovakia Slovenia Lithuania Cyprus Latvia Estonia Malta

EMU-25 – ranked by share in ECB capital Germany UK France Italy Spain Poland Netherlands Belgium Sweden Austria Greece Portugal Denmark Czech Rep. Hungary Finland Ireland Slovakia Lithuania Slovenia Latvia Luxembourg Estonia Cyprus Malta

Note: Classification is based on estimated 2002 data (2001 data in the last column). NMS grow, on average, faster than OMS so one can expect gradual changes in the ranking. Source: Gros (2003) and Lommatzsch and Tober (2002).

Yet another important criticism regards the criteria for the ranking of MS and consequently the division into groups. The inclusion of the indicator of the size of the financial sector (TABS-MFIs) with a weight of 1/6 on top of GDP (weight 5/6), without taking into account population shares must be viewed as arbitrary. There seems to be no convincing argument for choosing such a set of indicators and ignoring the benchmark provided by each national CB’s shares in the ECB capital that is applied in many decision-making procedures in the Bank15. It is hard to avoid the conclusion that this set of indicators was tailored to favor the interests of Luxembourg (a very small economy with a large financial sector) and, more generally, of the OMS at the expense of the NMS (whose share in EMU aggregate GDP and TABS-MFIs is much below their share in population). As illustrated by Gros (2003), the adoption of these criteria results in the third rotating group being composed of the NMS only and leaves Poland out of the Big 5 group even if the UK chooses not to enter the EMU before Poland does (see Table 11.2). Also, estimations based on 2002 data would put Luxembourg (population 0.45 million) in Group 2, while Romania, with a population of 21.7 million, would be in Group 3. On the

196

WOJCIECH PACZYNSKI

other hand, Table 11.2 also reveals that adopting the shares in the ECB capital as a basis for the ranking would not change the allocation to groups radically. Still, this system is likely to be perceived as unfair in NMS, particularly as these countries had no say on the reform. This could pose risks for the democratic legitimacy of the ECB. 8. CONCLUSIONS There is very little one can say with certainty on the optimal design of the monetary policy decision-making processes in a large and heterogeneous MU such as the EMU. This study argues that EMU enlargement is unlikely to undermine the effectiveness of the current rules. In particular, it claims that the oft-referred to “small countries’ bias” resulting from regional policy preferences among members of the GC is unlikely to distort ECB decisions, provided there are at least a few members of the GC taking a Union-wide perspective. The evidence presented indicates that several different solutions for ECB functioning might work well, despite their potential shortcomings. The currently applied one-man one-vote, simple majority procedure appears to be performing well, if judged from numerous reports monitoring the ECB’s performance in its early years. The reform proposed by the ECB and adopted by the EU Council also provides a reasonable framework for monetary policy-making in the larger EMU. The main weakness of both the old and the new proposed rule is related to the size of the decision-making body, which would complicate discussion prior to voting on interest rates. More generally, the major drawbacks of the adopted reform proposal are rooted in the sphere of intra-EU politics, rather than in economic effectiveness. Discussion on the ECB’s functioning will certainly continue in years to come given that the new rule will not become operational before 2008-2009. An important policy question is whether another change of voting modalities in the ECB is likely before that time. The design of the current ECB architecture was influenced by the experience of the functioning of the Bundesbank. Since the UK has stayed outside the Eurozone, the practice of direct inflation targeting frameworks steered by a small MPC consisting of monetary policy experts has had much less impact16. Such ‘policy culture’ influences arguably have a strong impact on institutional design. Consequently, one should not expect a major revolution in the decision-making rules of the ECB in the near future.

NOTES 1

The term ‘Member States’ applies here only to countries that have adopted the common currency. Only representatives of these governments appoint the Executive Board members (cf. EC Treaty, Art. 122 (4) and 112 (1)). While at present unanimity is required for appointment decisions, the Constitutional Treaty (Art. III-382) introduces qualified majority voting in this instance. 2 Also, the Treaties do not explicitly demand that CB governors of member states take into account unionwide developments instead of developments in their home economies. Governors are only banned from seeking advice or taking orders from EU or national bodies.

ECB DECISION-MAKING IN AN ENLARGED EMU

197

3

See ECB (2003) and references therein. Gruner & Kiel (2004) discuss the issue in more detail and derive some theoretical implications. 5 After German unification, a reform was implemented that amounted to the merger of some regional banks and reduced the size of the Council to 15 people. 6 The expected increase of the size of the Council was one of the reasons for the Bundesbank reform at the time of unification. 7 No information is available on the GC discussions and voting. This assertion rests on some public comments made by GC members. 8 This stands in contrast to the situation in the EU Council of Ministers, where the threshold for majority voting is well above 50%. 9 Note that there are above 180 countries in the IMF so the scale of the ‘numbers problem’ is of a different nature. Also, for most countries, decisions taken by this body are of much lower relevance in normal economic times than is the case of GC decisions in the EMU. 10 ECB capital shares are in turn determined as the sum of each country’s share in EMU GDP and of its share in EMU population, equally weighted (cf. Article 29, Protocol on the Statute of the ESCB and of the ECB). 11 See Aksoy, De Grauwe & Dewachteret (2002) for a discussion of how the weighted average of countries’ optimal policies corresponds to EMU optimal policy. 12 Alan Greenspan, December 2000. Cited in Meade & Sheets (2004). 13 This Decision of the Council was published in the Official Journal of the European Union, L83, Vol. 46 (1), April 2003, 66-68. 14 It is instructive to compare this with the QMV decision-making rule for the European Council agreed in Nice that allocates voting rights more fairly to MS, but sets the QM threshold at a very high level. 15 For instance all GC decisions concerning issues such as the capital of the ECB, policies with regard to foreign reserve assets, allocation of profits and losses of the ECB, etc. are taken by QMV with weights equal to each country’s shares in ECB capital, and the QM threshold set at two thirds (cf. ECBS Statute, Article 10.3). Shares in the ECB capital are equal to the sum of half of the share of a respective MS in the union’s population and half of its share in the union’s GDP (cf. ECBS Statute, Article 29). 16 Begg et al. (2003) argue that countries staying out of particular EU arrangements have limited impact at the stage of building respective institutions. 4

REFERENCES Aksoy, Y., De Grauwe, P. & Dewachter, H. (2002). Do Asymmetries Matter for European Monetary Policy? European Economic Review, 46 (3), 443-469. Angeloni, I. & Ehrmann, M. (2004). Euro Area Inflation Differentials. ECB Working Paper, 388. Baldwin, Richard E., Erik Berglof, Francesco Giavazzi & Mika Widgren. (2001). Eastern enlargement and ECB reform. Swedish Economic Policy Review 8, 15-50. Begg, D., Blanchard, O., Coyle, D., Eichengreen, B., Frankel, J., Giavazzi, F., Portes, R., Seabright, P., Venables, A., Winters, L.A., & Wyplosz, C. (2003). The consequences of saying no: An independent report into the economic consequences of the UK saying no to the Euro. Available at http://195.157.85.85/Begg.pdf. Black, D. (1958). Theory of Committees and Elections. Cambridge. Eichengreen, B. & Ghironi, F. (2001). EMU and Enlargement. Mimeo, available at http://emlab.berkeley.edu/users/eichengr/research/fabiobrussels10.pdf. ECB (2003), Inflation Differentials in the Euro Area: Potential Causes and Policy Implications. European Central Bank. De Grauwe, P. & Senegas, M-A. (2003). Monetary Policy in EMU When the Transmission is Asymmetric and Uncertain. CESifo Working Paper, 891. Gerling, K., Gruner, H.P., Kiel, A. & Schulte E. (2003). Information Acquisition and Decision Making in Committees: a Survey. European Central Bank Working Paper, 256. Gros, D. (2003). Reforming the Composition of the ECB Governing Council in View of Enlarged: an Opportunity Missed. CEPS Policy Brief, 32. Gruner, H.P. & Kiel, A. (2004). Collective decisions with interdependent valuations. European Economic Review, 48, 1147-1168.

198

WOJCIECH PACZYNSKI

Heisenberg, D. (2003). Cutting the Bank Down to Size: Efficient and Legitimate Decision-making in the European Central Bank After Enlargement. Journal of Common Market Studies, 41 (3), 397-420. Honohan, P. & Lane P.R. (2003). Divergent Inflation Rates in EMU. Economic Policy, 18 (37), 357-394. Lommatzsch, K. & Tober, S. (2002). Monetary Policy Aspects of the Enlargement of the Euro Area. Deutsche Bank Research Notes, 4. Maier, P. & Cavelaars, P. (2003). Convergence of Price Levels: Lessons from the German Reunification. Economics Working Paper Archive at WUSTL, Macroeconomics, 0306016. Meade, E.E. & Sheets, D.N. (2004). Regional Influences on FOMC Voting patterns. Journal of Money Credit and Banking, forthcoming. Myerson, R.B. (1996). Fundamentals of Social Choice Theory. Discussion Paper, 1162. Center for Mathematical Studies in Economics and Management Science, Northwestern University.

MAREK DABROWSKI

CHAPTER 12

A STRATEGY FOR EMU ENLARGEMENT

1. INTRODUCTION May 1, 2004 brought the biggest enlargement in the history of the European Union. Ten countries, eight of them from the former Soviet block, were admitted as new member states (NMS) to the EU. However, despite its symbolic character, this date neither began nor ended the integration process. All the post-communist NMS have already gone through a series of deep, radical, and sometimes painful reforms for the last 15 years, which allowed them to build the foundations of democratic capitalism and prepare them for EU membership. This process was accompanied and supported by the series of Trade and Association agreements signed between the EU and candidate countries and among some of the candidate countries. Six years of accession negotiations brought further economic and institutional adjustment. As a result, NMS are at present much more closely integrated with the old member states (OMS) than was the case with Greece, Spain or Portugal when they entered the EU in the 1980s. However, the date of formal EU entry did not complete the accession process. One of the remaining tasks is joining the Economic and Monetary Union (EMU). Formally, NMS do not have an opt-out clause like the UK or Denmark, and as such they are expected to join the EMU soon. However, the Accession Treaty does not contain a timetable for entry, which will formally depend on states meeting the Maastricht criteria of nominal convergence. In practice, this leaves significant room for maneuver in setting the date for joining the eurozone, as confirmed by the example of Sweden. In both the NMS and the OMS, politicians, economists and public opinion are divided on the optimal timing of NMS joining the eurozone. The advocates of fast EMU entry stress the high level of trade and business cycle integration of NMS with the eurozone, and the potential benefits for NMS in terms of eliminating transaction costs and exchange rate risk. Opponents of fast eurozone entry underline the costs of meeting the Maastricht criteria and giving up the supposed shock-absorbing role of the exchange rate. In addition, in OMS there are some political and economic concerns. The former come down to retaining a carrot which can be granted or withheld from NMS depending on their good behavior, and some understandable concerns about how responsibly they are likely to behave after EU accession. 199 M. Dabrowski and J. Rostowski (eds.), The Eastern Enlargement of the Eurozone, 199-225. © 2006 Springer. Printed in the Netherlands.

200

MAREK DABROWSKI

Economic fears relate mainly to the controversial hypothesis that accession of fast growing countries will increase inflationary pressure and interest rates in the eurozone, which will have an additional contractionary impact on the slower growing economies of some OMS. Another controversy relates to the path leading to EMU accession and in particular to the Exchange Rate Mechanism (ERM-II), which needs to be adjusted to the new realities of international financial markets dominated by unrestricted capital movements. This chapter tries to address these dilemmas and propose policy recommendations both for NMS (on how to manage their accession to the eurozone) and for the European Commission, ECB and OMS (on how to manage and absorb EMU enlargement in an optimal way). It also summarizes the findings of other chapters in this volume. Section 12.2 starts with a review of arguments about EMU accession by NMS, based on optimum currency area (OCA) theory, both in its classical (static) and modified (dynamic) versions. Section 12.3 confronts the argument that rapid nominal convergence may be harmful for real convergence. Section 12.4 discusses the formal criteria and procedures of EMU membership and EMU accession as set by the Treaty, and assesses the room of maneuver available for various EMU accession strategies. Section 12.5 analyzes how the EMU will function after its enlargement. Finally, Section 12.6 contains conclusions and recommendations. 2. WHAT CAN BE LEARNT FROM OCA THEORY? Discussions of the advantages and disadvantages of membership in monetary unions (MU) usually start from OCA theory and try to find empirical evidence verifying the arguments of the theory in relation to a particular country or group of countries. 2.1. Original OCA theory and its modification The original OCA theory in the version proposed by Mundell (1961) and McKinnon (1963) concentrates on so-called asymmetric shocks, i.e. shocks affecting two territorial units in a different way. In a world of sticky wages and prices there are basically three possibilities of addressing the negative social consequences of such shocks, i.e. (1) move production factors (labor, capital) between regions, (2) use fiscal transfers to help a disadvantaged region and (3) change the exchange rate between these regions. If we cannot count on (1) or (2) the exchange rate remains the only available adjustment tool. So a common currency for two territories makes sense only if the risk of asymmetric shocks is small, there is factor mobility or the possibility of fiscal redistribution between them. On the other hand, multiplicity of currencies and floating exchange rates increase transaction costs and harm trade. Although Mundell (1961) was very careful in balancing potential advantages and disadvantages of exchange rate flexibility and did not formulate any strong and explicit policy recommendations, his paper could be interpreted as rather discouraging for the idea of introducing a common currency for territories not being

A STRATEGY FOR EMU ENLARGEMENT

201

part of a political union, because only the latter allowed for deeper fiscal redistribution and made factor mobility easier. Furthermore, exchange rate adjustment could be considered as a convenient and rather costless way of stabilizing output and employment in response to adverse shocks. This was slightly corrected by McKinnon (1963) who stressed that currency depreciation in a small open economy must inevitably lead to higher inflation. Generally speaking, the original OCA theory reflected the post-WWII economic landscape with its extensive trade, current account and capital account restrictions, its increasing microeconomic rigidities and an almost universally activist approach to monetary and exchange rate policies aimed at stabilizing output and employment rather than fighting inflation (see McKinnon, 2000). OCA theory evolved in the subsequent decades, partly as a result of serious changes in world economic and financial architecture. Global product and financial markets have become increasingly integrated. Thus, international factor immobility, one of the key assumptions of Mundell’s (1961) model, has partly disappeared. In his later papers Mundell (1973a; 1973b) became more active in demonstrating the advantages of a common currency, particularly in the case of European Economic Community, discussing the first blueprints for EMU. He raised three new arguments in favor of monetary unification – risk pooling, saving international reserves due to ‘internalization’ of the formerly foreign trade, and benefits of scale connected with lowering the costs of financial transactions – which significantly changed his original balance of costs and benefits. 2.2. Endogeneity of OCA criteria and political determinants of MU Another weakness of the original version of OCA was due to its static character. Asymmetric shocks and limitations in factor mobility were viewed as exogenous. However, monetary unification will promote trade and capital flows within the common-currency area, and create pressures to synchronize fiscal policies. It may also encourage the further deregulation of labor, capital and product markets. All these factors will reduce the incidence of asymmetric shocks and increase factor mobility. In fact, already Mundell (1961) noted this problem, quoting an earlier view of Scitovsky (1958) who believed that a common currency in Western Europe would induce a greater degree of capital mobility. However, it was Frankel & Rose (1998) and Rose (2000) who analyzed the question of the endogeneity of OCA criteria in depth and supported it by a large-scale cross-country analysis of the influence of a common currency on trade creation and synchronization of business cycles. Looking historically, most national currency areas have had a political origin and design (formation of national states, colonial empires, political unions) and both trade flows and factor mobility have had to adjust (and did adjust in most cases) to politically determined borders.

202

MAREK DABROWSKI

2.3. Costs and benefits of sovereign monetary policy Finally, one may ask what the real benefits of exchange rate flexibility are, as a tool for the stabilization of output and employment in a world of free capital movement, particularly in the case of small economies which trade a great deal? Under a fully floating exchange rate regime nominal (and real) exchange rates may automatically adjust to changing terms of trade and changes in capital flows unless other factors push in the opposite direction. As regards discretionary policy moves, authorities can change nominal exchange rates (NER) but this does not mean that they are able to achieve any durable change in the real exchange rate (RER). For example, a discretionary currency devaluation aimed in boosting external competitiveness will most likely end with higher inflation eating up the real effects of the nominal devaluation. Flexible exchange rates also leave some room for discretionary monetary policies focused on achieving goals other than price stability. However, giving up monetary independence does not fully eliminate the counter-cyclical role of monetary policy. If the central bank of a common currency area does pursue such a policy and the business cycle conditions do not differ significantly between MU members, monetary conditions will help to close any output gap. The problem occurs when a country must deal with asymmetric shocks (i.e. shocks not affecting other members of a MU or affecting them differently), notwithstanding the influence of monetary unification in harmonizing business cycles and promoting trade inside a common currency area. There is, however, another question: to what extent is the central bank in a small open economy able to ‘lean against the wind’ – particularly if we take into account the issue of credibility (see below) and political and technical constraints. Any monetary policy discretion needs to be backed by a strong conceptual and political anti-inflationary consensus, which is not always present in less developed economies, including NMS. Second, discretionary monetary policy (for example, the direct inflation targeting adopted in the Czech Republic and Poland) needs welldeveloped analytical and forecasting skills, both inside and outside the central bank, as well as high-quality monetary, fiscal, price, balance of payments and output statistics. If a central bank wants to build its credibility in order to achieve rapid disinflation at relatively low output/ employment costs, it must stick quite resolutely to its low inflation target, without compromising it in favor of other goals. This means giving up most of the potential advantages that an independent monetary policy is supposed to have compared to a lack of monetary sovereignty. 2.4. Criteria of empirical verification The following indicators are most frequently used to assess how well prepared NMS are to become a part of the EMU: 1. trade links with the EU-15 or with the current eurozone, including the role of intra-industry trade, 2. synchronization of business cycles,

A STRATEGY FOR EMU ENLARGEMENT

203

3. actual exchange rate variability (nominal and real), 4. actual factor mobility between the NMS concerned and OMS The main rationale behind the first two indicators is the belief that the deeper trade integration and the closer co-movement of business cycles between countries, the smaller is the probability of idiosyncratic shocks affecting them. Furthermore, close trade links increase the benefits of a common currency in terms of lower transaction costs and stable prices. The third measure is supposed to provide a historical picture of how frequently and how deeply exchange rates had to adjust, indicating the frequency of various shocks affecting the economy, or at least its balance of payments. However, one must remember that exchange rate variability may reflect incidence of both real and nominal shocks. The latter are very frequently caused by a national monetary policy. Finally, the fourth type of empirical test should tell us to what extent international factor mobility can help in cushioning the consequences of asymmetric shocks. The incumbent EMU members, or their less developed Mediterranean sub-group, usually serve as a benchmark for comparison. Such an approach implicitly assumes that the current eurozone is an OCA which is not necessarily the case1. 2.5. Trade integration and synchronization of business cycles Both Rostowski (Chapter 1 of this volume) and Blaszkiewicz-Schwartzman & Wozniak (Chapter 3) demonstrate in their empirical analyses the very high level of trade integration of NMS with OMS which was achieved already in the late 1990s/ early 2000s. This level was similar, or even higher, than that of incumbent members at the time. It is worth stressing that both studies do not take account of trade between NMS themselves. Assuming that all NMS will eventually enter the EMU, this means that their trade exposure to other members of the common currency area will be even higher2. As regards the synchronization of business cycles, the results of the empirical work3 discussed in Chapters 1 and 3 are somewhat ambiguous. BlaszkiewiczSchwartzman & Wozniak (Chapter 3) develop their own empirical analysis of business cycle correlation. They examine correlations between changes in real GDP, industrial output and employment between the eurozone and individual countries (both eurozone members and NMS) for the period 1994-2002, as well as separate analyses for shorter sub-periods, based on both annual and quarterly data. Unlike in the case of exports to EU/ EMU, the picture obtained by them in the area of business cycle correlations seems to be less optimistic for the perspective of NMS membership in the eurozone. With the exception of Hungary and Slovenia, other NMS record weak or even negative correlations of their business activity parameters with those of the eurozone (in particular as regards the unemployment rate). So, NMS might be, on average, more exposed to a danger of idiosyncratic shocks than the current EMU members. There are three caveats, however. First, from purely statistical point of view correlation coefficients for eurozone members include a certain component of autocorrelation, particularly strong for the biggest countries (because the growth and

204

MAREK DABROWSKI

unemployment dynamics for the eurozone have been calculated as a weighted average of its members. Second, NMS went through a deep and painful restructuring in the 1990s and at the beginning of the 2000s, as part of their post-communist transition and the structural/ institutional harmonization of their economies with those of EU members. Neither the transition-induced output decline nor the ensuing recovery have followed business cycle patterns in the EU economies. In addition, changes in employment did not fully follow changes in output dynamics because, again, of deep restructuring and, to some extent, because of labor market rigidities (see Chapter 5). Furthermore, it is very difficult to find evidence of any business cycle (in its classical meaning) in most of transition economies, at least until the end of 1990s. Finally, the endogeneity hypothesis should be also taken into consideration, at least partly. Some of the current EMU members fixed the exchange rates of their currencies to the German Mark well before they formally joined the ERM-I and EMU. This relates, above all to Austria and the Benelux countries, which have effectively belonged to a D-Mark zone at least from the beginning of 1980s. On the other hand, weak or negative correlation coefficients of Lithuania and Latvia are hardly surprising as the former had a dollar-denominated currency board between 1994 and 2001 (and USD recorded substantial fluctuations vis a vis DEM and EUR during this period) and the latter pegged its currency to SDR. All three arguments call for care in gauging the future appropriateness of individual countries becoming members of EMU based on their historical record, and reflect the logic of the Lucas critique. Exploring further the question of endogeneity, Maliszewska (Chapter 4) assess the potential for additional trade expansion after the NMS will join the EMU. In her simulation, based on a gravity model, NMS trade with eurozone countries (and among themselves, assuming that all they will join the EMU at approximately the same time) should expand beyond what can be expected as result of their accession to the EU (see Maliszewska, 2004). The biggest potential gains should be enjoyed by those NMS (Latvia, Lithuania and Poland), which have not yet reached the level of trade integration typical for members of the EU. Trade expansion between NMS following euro adoption should be even higher, particularly for Slovenia and Poland. The smallest gains may be recorded by the Czech Republic and Slovakia because of their high level of bilateral trade, reflecting the legacy of the Czechoslovak federation. The results of another empirical study (Brzozowski, 2003) point to the possibility of bigger FDI inflow to NMS after their EMU accession. Both increased intra-EMU trade and FDI flows should lead to a closer correlation of business cycles between NMS and the EMU, decreasing the probability of asymmetric shocks. 2.6. Exchange rate variability The examination of exchange rate variability carried out in Chapter 3 follows the approach adopted by Vaubel (1976; 1978), von Hagen & Neumann (1994) and Gros & Hobza (2003). Authors show that according to this criterion NMS in the early

A STRATEGY FOR EMU ENLARGEMENT

205

2000s have been, on average, similar to the Mediterranean countries in the early, rather than mid 1990s. Two countries – Estonia and Slovenia – have exhibited fluctuations of RER similar to or lower than the Club Med countries (both for a more turbulent period - 1993-1995 - and a less turbulent one - 1996-1998). The RER variability of Bulgaria, the Czech Republic, Hungary, Latvia, Lithuania, Poland, Romania and Slovakia has exceeded that of Mediterranean countries. Two important comments must be made at this point. First, even countries, which do not record NER fluctuations (because of their currency boards or ERM-I membership) are subject of RER changes, which means that there are other ways of accommodating shocks like inflation differentials or differences in productivity gains. Second, we should raise again questions of endogeneity and causality. As we mentioned before, apart from real shocks such as, for example, a dramatic change in oil prices, national economies are exposed to nominal shocks caused by domestic macroeconomic policies, including monetary and exchange rate shocks. Thus, it should not be surprising that in Blaszkiewicz-Schwartzman and Wozniak study (Chapter 3), Estonia and Slovenia recorded the lowest RER variability. Estonia gave up any sovereign monetary policy in 1992, introducing a currency board and pegging its currency (the Kroon) to the German Mark and then the euro. Slovenia followed an informal crawling peg to the D-Mark and then the euro from 1991, which also seriously limited the room for a discretionary monetary policy. When Mediterranean countries finally locked the exchange rates of their currencies in the ERM-I system in the second half of 1990s, variability of their RER also went down. Maliszewska & Maliszewski (Chapter 2) investigate a much bigger sample of 144 countries for a period of more than 60 years (1940-2001) and find that (i) hard pegs help to improve inflation performance; (ii) de facto pegged regimes as defined according to the ‘natural’ classification (see Reinhard & Rogoff, 2002) are associated with lower output volatility. Similar conclusions can be drawn from Kowalski’s (2003) empirical research covering seven NMS. Fixed exchange rates helped these countries in achieving both better inflation performance and higher growth rates. Interpretation of these findings can go in one of two directions: x Floating exchange rates do not necessarily help to cushion shocks. On the contrary, they can cause greater volatility and instability of both nominal and real variables. Thus, they often play the role of shock generator instead of shock absorber. x In the medium to long run, the benefits of exchange rate stability and low inflation can outweigh risks connected with giving up national monetary policy. We will come back to this issue in the next section. 2.7. Factor mobility Regarding factor mobility, one should note that NMS have completely opened their economies to free capital movement. The Single European Market involves free movement of labor between member states (MS)4, although the actual cross-country

MAREK DABROWSKI

206

labor mobility is far from being perfect. However, internal labor mobility (particularly geographical) within most EU countries is also highly imperfect due to numerous labor market rigidities (see Chapter 5). So the actual EU picture (including NMS) is far from that originally assumed by Mundell (1961), i.e. full domestic mobility and full external immobility of production factors. Empirical research presented in Chapter 5 shows that, generally, nominal wages in NMS are not downwardly flexible, although the degree of rigidity differs between countries. Lithuania is the only exception, demonstrating some nominal flexibility in its wage setting. As this country has run a currency board from 1994, this may point again to the endogeneity of OCA criteria. Fast productivity growth in NMS is another mechanism, which allows for adjustment even when nominal wages cannot be easily cut (see Chapter 5). Whether introducing the euro and ultimate giving up the possibility of competitive devaluation will further boost productivity growth remains an open question. 2.8. Summary Summing up this part of our analysis, NMS seem to be relatively well prepared for EMU membership. Taking into consideration the ‘classical’ OCA criteria, i.e. trade integration, co-movement of the business cycles and actual factor mobility with the rest of the currency area their record is not worse, on average, than that of current eurozone members at a similar stage of their EMU accession process. Assuming a continuation of rapid integration with OMS and among themselves, NMS can further decrease their exposure to idiosyncratic shocks before entering the eurozone. After joining the EMU, monetary unification should help them to develop additional intra-EMU trade links, further synchronize business cycles and increase factor mobility with the rest of the MU (endogeneity hypothesis). 3. IS THERE A TRADE-OFF BETWEEN NOMINAL AND REAL CONVERGENCE? Both Western and Eastern skeptics regarding fast EMU enlargement argue that while NMS joining the eurozone can speed up their nominal convergence with the rest of the EU, it will not necessarily help them with real convergence. It is claimed, therefore, that they should make progress in real convergence first and only then think about EMU membership, which requires nominal convergence and tough macroeconomic policy to meet the Maastricht criteria. As with the debate discussed in the previous section, the key question relates to the role of discretionary monetary and fiscal policies: are they helpful tools in speeding up the long term rate of economic growth? Or, putting it in another way, can the macroeconomic discipline, required by the Treaty on European Union and the Stability and Growth Pact (SGP), be harmful for economic growth?

A STRATEGY FOR EMU ENLARGEMENT

207

3.1. Definitional problems While the debate on OCA criteria is well anchored in the economic theory this is not the case with the controversy on nominal vs. real convergence. The difficulties start already with understanding what the two types of convergence precisely mean. There are fewer problems with defining nominal convergence. It is broadly understood as ‘…convergence of certain macroeconomic indicators to levels ensuring macroeconomic stability in an economically integrated area’ (Kowalski, 2003).

Or more concretely, EMU candidates approaching the numerical values of inflation and long-term interest rates prevailing in the eurozone, stabilizing their exchange rates vis a vis the euro during the two years of ERM2 membership, and respecting fiscal discipline as defined by the Treaty (see Section 12.4.2). The meaning of a real convergence is not so clear. One may find at least two interpretations of what ‘real convergence’ means. The first, which dominates in the economic literature on growth and ‘catch-up’, refers to diminishing the welfare (income, living standard, productivity, etc.) gap between poor and rich countries (regions) as result of faster growth of the former. This meaning of real convergence is very close to the term ‘cohesion’ used frequently in the official EU jargon. Economic cohesion involves a set of policies at the EU level aiming at a reduction of income disparities between regions and MS (Ardy, Begg, Schelke & Torres, 2002)5. A second interpretation of ‘real convergence’ can be indirectly found in the European Commission convergence reports (see CR, 2004a; 2004b). These documents speak generally about ‘convergence’ and do not distinguish between the ‘real’ and ‘nominal’. However, the content of CR allows one to conclude, that apart from nominal and legal criteria the Commission also analyzes a set of criteria related to the real economy, like financial and product market integration, development of the balance of payments, FDI and trade flows, unit labor costs and other price indices6. According to CR (2004b), they are ‘…relevant to economic integration and convergence (p.6) [and] …when the economic and monetary union is hit by a shock, a high degree of convergence limits the emergence of asymmetric economic developments at the country level, to which not any longer can be responded by using the exchange rate’ (p.1).

So from the overall context of the CR one can conclude that real convergence is understood as integration of national product and financial markets7 of EMU candidates with those of the remaining EU/ EMU members rather than accomplishing any numerical targets of income catching up. In fact this kind of understanding of real convergence corresponds quite closely to the OCA theory discussed in Section 12.2. And if one believes that progress in real convergence is to determine country eligibility for joining a MU only the second kind of interpretation makes sense. There is no theoretical foundation to claim that countries or territories with substantial differences in level of economic development cannot share a common currency. As one of the arguments behind the EMU project was exactly a belief that it would speed up economic growth and

208

MAREK DABROWSKI

reducing the income disparities of its members (see ‘Delors Report’, 1989; Emerson et al., 1992; Jarocinski, 2003) it would be strange to reverse this causality in relation to the current EMU candidates, and expect that they will first catch up and only then become eligible to join the eurozone. In practical terms, as differences in development level between most of NMS and OMS are quite substantial now and closing this gap will require a lot of time (see e.g., Fischer, Sahay & Vegh, 1998), this approach would delay the perspective of EMU enlargement for many years if not decades (see Chapter 1). Still, questions remain: first, to what extent will EMU membership help in achieving nominal and real convergence (the later understood as diminishing the development gap); and second, will fast nominal convergence be supportive of real convergence? We examine each of these questions, both from a theoretical and an empirical perspective. 3.2. MU and nominal convergence Numerous empirical observations demonstrate that a credible exchange rate anchor (in the form of currency board) or giving up the national currency in favor of dollar or euro may be very helpful in achieving sustainable disinflation, price stability and either elimination or substantial reduction of interest rate spreads vis a vis an anchor currency (due to elimination of exchange rate risk). Such positive effects are especially evident in less developed and transition countries, which face huge problems in building a wide political consensus around the goal of price stability, lack a credible track record for their central banks, which may be technically incapable of running a really independent and price-stability-oriented monetary policy (see Section 12.2.3). Even if sovereign monetary policy manages to achieve price stability in such circumstances it usually comes at substantial cost in terms of high interest rates and nominal appreciation of the exchange rate and strong exchange rate fluctuations (see Chapter 1). Looking at the five so-called Maastricht criteria of nominal convergence, credible exchange rate stability has primary importance being, in fact, exogenous in relation to both the inflation and (particularly) the interest rate criterion. In other words, a country that decides to credibly peg its currency to the euro forever enjoys a rapid convergence of interest rates and – if its inflation level significantly exceeds that of eurozone – also convergence of the inflation level. Evidence of how a fixed exchange rate can effectively influence the disinflation process and interest rate convergence can be found in numerous empirical studies (see Chapters 2 & 6; Kowalski, 2003). In his comparative study analyzing the EMU accession experiences of Greece, Portugal and Spain, Jarocinski (2003) demonstrates that entering the ERM and the serious fiscal tightening required by the Maastricht criteria led to rapid disinflation and interest convergence in these countries. However, after the launch of EMU in 1999, some of its members (Ireland in the first instance but also Spain, Greece, Portugal, the Netherlands, Italy) started to experience significantly higher inflation than the average of the eurozone (see Jarocinski, 2003). This could be explained by several factors, such as higher rate of

A STRATEGY FOR EMU ENLARGEMENT

209

productivity growth and its consequences for price structure (the so-called Harrod – Balassa – Samuelson [HBS] effect), rapid depreciation of EUR in relation to USD and GBP, which increased inflationary pressure in countries dependent on trade denominated in these currencies (e.g. Ireland – see Honohan & Lane, 2004), original price level disparities, particularly observed in lower income EMU members, and fiscal expansion. The causal link between entering a MU and the strengthening of fiscal discipline is not as obvious as in the case of inflation and interest rate convergence. On the one hand, giving up monetary sovereignty and with it the possibility of corrective devaluation, ties policymakers’ hands and forces them into fiscal prudence. Furthermore, interest rate convergence can bring huge fiscal gains, making fiscal adjustment much easier. Such benefits could be observed in particular in the process of EMU accession by highly indebted countries suffering credibility problems (Italy and Greece). The same may happen in the case of relatively highly indebted and high-interest-rate NMS like Hungary or Poland (see Gorzelak, 2004). On the other hand, a rapid decrease of interest payments creates a temptation to substitute them with other spending programs, instead of using this opportunity for improving the fiscal balance. Worse, there is a risk of free riding behavior when government, exploiting the credibility of a MU, can raise debt at relatively low cost and without the danger of being quickly punished by financial markets (see Chapter 9). This is the main rationale behind the two fiscal criteria in the Maastricht Treaty and the additional fiscal discipline rules of the SGP (see Section 12.4.2). 3.3. Nominal convergence and growth prospects Let us turn to the issue of how nominal convergence (speeded up by joining the EMU) can influence the growth prospects of NMS. The potential costs and benefits of giving up monetary sovereignty and flexible exchange rate have been already discussed in Section 12.2. A positive impact of lower real and nominal interest rates seems to be also beyond controversy and does not require in further elaboration. The remaining question is whether lower and more stable inflation and bigger fiscal discipline can be helpful or harmful for economic growth, particularly in middleincome transition countries? There is a vast economic literature giving a positive answer to both parts of this question. Low inflation is positively correlated with economic growth in the medium to long run8 and the same is true for fiscal prudence. The main benefits coming from stable prices and balanced budgets are of two kinds. First, the rate of private savings should be higher, other things being equal, and smaller portion of savings would be eaten up by the public sector borrowing requirements (the crowding out effect). Second, in a zero or very low inflation environment market forces should work better and both prices and wages should be more flexible (one should remember that sticky prices and wages are the main argument in favor of exchange rate flexibility). Nevertheless, in the short run both disinflation and fiscal adjustment may involve output and employment losses. While this argument cannot play any substantial role in a discussion of real convergence (where only the long-term perspective matters) it

210

MAREK DABROWSKI

is a real dilemma for many politicians and policymakers, whose preferences may be determined by the proximity of elections (see Alesina and Perotti, 1994). There are three additional circumstances, however, which make the short-term policy trade off less dramatic, at least for NMS. First, there is an increasing number of empirical cases in which discretionary fiscal policy brings effects quite opposite to those expected by advocates of fiscal activism, i.e. higher growth results from fiscal tightening and lower growth is a consequence of fiscal expansion (see Chapter 8 for broad empirical evidence). Such ‘non-Keynesian’ effects of fiscal policy seem to be determined by the expectations of economic agents and their saving behavior. If government is already close to its borrowing limit (determined by financial markets’ judgment on the actual and expected size of deficits and debt or formal fiscal rules like those existing in the EU) they know that fiscal adjustment (in the form of higher taxes or lower expenditures) will come soon and they therefore increase precautionary savings, decreasing current demand. The opposite will then be true in the case of fiscal consolidation (see Bertola & Drazen, 1993; Sutherland, 1997; Chapter 8). Second, all MS of the EU are obliged to respect the fiscal rules in the Treaty and the SGP, irrespective of their participation in the EMU. Thus, those NMS, which postpone their EMU accession, will have to follow fiscal discipline rules anyway and bear all the associated costs of fiscal consolidation, without enjoying benefits of joining a common currency. Finally, as regards inflation, most of NMS have already brought inflation down and paid the costs of this process, and the question now is how to stabilize low inflation in a credible way. Here joining the eurozone offers a much better perspective than any national monetary fine-tuning. Regarding empirical evidence on the positive correlation between nominal and real convergence, this has been provided, among others, by Kowalski (2003), Jarocinski (2003) and Chapter 2. Summing up, it is hard to find any logical argument or empirical evidence that entering the EMU and the nominal convergence associated with it can hurt the economic growth of NMS. On the contrary, EMU entry will likely accelerate catchup in the long run, helping real convergence or economic cohesion9. 4. THE ROCKY ROAD TO EMU The NMS appear to be relatively well prepared to join the eurozone (according to the OCA criteria) and it should help them both in nominal and real convergence – these are the main conclusions from the two previous sections. So what should be done in order to achieve this policy goal? 4.1. Member States ‘with derogation’ According to Article 4 of the Treaty of Accession (signed in April 2003 in Athens) NMS obtained the status of ‘Member States with derogation’ regarding the EMU membership (CR, 2004b, p. 2). Although they do not have formally an opt-out

A STRATEGY FOR EMU ENLARGEMENT

211

option like Denmark and UK, a concrete date for their accession depends on their meeting the nominal convergence and legal criteria of the Treaty. This provision gives the NMS great room for maneuver as to when they will adopt the common currency. On the other hand, it is also hypothetically possible to postpone the EMU accession of NMS by such an interpretation of entry criteria by the EMU side that they become difficult to fulfill or just through a failure by the European Council to decide to admit the new EMU member(s)10. 4.2. Nominal and legal criteria and the traps associated with them The four nominal convergence criteria as defined by the Treaty are as follows: x average inflation rate over a period of one year before the examination, that does not exceed by more than 1.5 percentage points that of, at most, the three best-performing MS in terms of price stability (Article 121(1) of the Treaty and Article 1 of the protocol on the convergence criteria); x exchange rate stability within the ‘normal fluctuation margin’ of the exchange rate mechanism (ERM-II) for at least two years ‘without severe tensions’; in particular, the EMU candidate must avoid devaluation of the central parity against the euro ‘on its own initiative’ (Article 121(1) of the Treaty and Article 3 of the protocol); x average long-term nominal interest rate over a period of one year before the examination, that does not exceed by more than 2 percentage points that of, at most, the three best-performing MS in terms of price stability (Article 121(1) of the Treaty and Article 4 of the protocol). x general government deficits below 3% of GDP (Article 121(1) in accordance with Articles 104(2) and 104 (6) of the Treaty) and general government debt level remaining below 60% of GDP or diminishing at a satisfactory pace towards the reference value (as above). In addition, EMU candidates must demonstrate compatibility of their national legislation, including the statutes of their national central banks, with Articles 108 and 109 of the Treaty and the Statute of the ESCB/ECB (see CR, 2004a; 2004b). Let us start with the legal criteria. The ‘Member States with derogation’ are expected to already have their legislation fully compatible with that of EMU members. Yet, surprisingly, none of the countries covered by the CR was assessed as meeting this requirement, in spite the fact that central bank legislation, including provisions related to central bank independence, was subject to Commission scrutiny already at the stage of EU accession negotiations, and had to be adjusted then. The very formal interpretation of the legal requirements in the last CR may signal the Commission’s intention to use them as an instrument to delay EMU enlargement. The four criteria of nominal convergence were formulated in the beginning of 1990s. Their purpose was to create a MU and new currency and not merely to enlarge its territory and number of members. Besides, the global macroeconomic and financial environment was quite different fifteen years ago when the

212

MAREK DABROWSKI

foundations of the EMU were designed. Some of EU-12 members still had capital controls and global financial markets were less developed and sophisticated. It is fair to say that from the very beginning mutual consistency of these criteria raised some doubts and in fact was possible only under some additional assumptions. The two requirements of the fiscal criterion represent the best example here (see Chapter 9 and Gros, Mayer & Ubide, 2004). The upper deficit ceiling of 3% of GDP is consistent with the upper debt ceiling of 60% of GDP only under the assumption of a 5% growth rate of nominal GDP. With a deficit of 3% of GDP and 2% annual inflation, real GDP must grow at least at 3% annually to avoid breaching the debt criterion in the longer. If average real growth is lower (which has been the case of the EU as a whole and of most of its members for a number of years) the deficit must be respectively lower. Countries growing faster than 3% per year or having slightly higher inflation (both cases can relate to NMS) could be allowed to run higher fiscal deficits without breaching the debt provision. Nevertheless, we do not recommend such a waiver because of: (i) the serious long-term fiscal risks facing most of EU members such as the negative consequences of population aging (Gros, 2004), already accumulated huge implicit public debt (Gomulka, 2001), numerous contingent fiscal liabilities (Polackova-Brixi, 2004) and others; (ii) uncertainty about future growth rates; (iii) general skepticism relating to the benefits that can be expected from higher deficits (see Section 12.3.3). A similar inconsistency affects the exchange rate stability and inflation criteria (see Dabrowski & Rostowski, 2001). Fixing the exchange rate makes the inflation rate mostly exogenous for the monetary authorities, particularly in a world of free capital movement. And inflation differences exceeding what is tolerable under the inflation criterion can result from various sources such as differences in productivity growth (HBS effect), changes in demand structure (in favor of non-tradable services), or initial differences in purchasing power parity (PPP) of individual currencies. NMS, growing faster than OMS and starting from lower levels of development can experience all these sources of higher inflationary pressure and thus face the danger of breaching the inflation criterion11. An additional difficulty can result from the fact that the reference value for the inflation criterion is calculated on the basis of a simple arithmetic average of the three best-performing EU members, which do not have to be the EMU members. With substantial differentiation of inflation rates inside the EU, it can happen that the three best performers representing a very small share of the overall EU GDP, can set the reference value well below the average inflation rate for the entire EU (and indeed EMU) and the actual rate of most of their members. Obviously, such a hypothetical (but not completely unrealistic) scenario does not make any sense from the point of view of nominal convergence of the candidates with the current EMU (they should adjust their inflation rates to the eurozone average rather than three EU25 outliers). What is the room of maneuver for the EMU candidates, particularly when they are already within the ERM-II and cannot use interest rates aggressively to bring inflation down (because this may cause excessive fluctuations of exchange rates)? First, fiscal and income policies may offer some help here (especially given that

A STRATEGY FOR EMU ENLARGEMENT

213

candidates will have to conduct very prudent fiscal policy anyway), but their potential impact may be limited. Second, there is the possibility of one-off revaluation of the exchange rate’s central parity, which does is not in conflict with the exchange rate criterion. Third, the central parity declared upon ERM-II entry should not be undervalued (see Section 12.4.4). Similarly, rules for the calculation of the interest rate criterion can also give random results, although here the risk of serious deviation of the reference value is smaller than in the case of inflation (because there is already far going interest rate convergence inside the eurozone and it is very unlikely that outsiders can significantly out-perform EMU members on interest rates). Here the danger is the opposite: outsiders with very low inflation rates may compose the reference group for the interest rate criterion, yet their long-term interest rates may be quite high. We next turn to the exchange rate criterion. 4.3. The ERM-II trial period The definition of the exchange rate stability criterion (see above) raises several practical questions. The first one concerns the meaning of the phrase ‘normal fluctuation margins’. After the 1992 ERM-I crisis this margin was extended to +/15% and this interpretation remains in force until now. In spite of the earlier suggestions of the European Commission and the ECB that a narrow corridor of +/2.25% around the central parity will be the preferred ERM-II mechanism, three NMS, which joined the ERM-II on June 28, 2004 (Estonia, Lithuania and Slovenia) are formally obliged to intervene if the exchange rate of their currencies is about to breach the wide +/- 15% band. However, another important question is how the wording ‘without severe tensions’ will be interpreted in practice. If the wide fluctuation band of +/- 15% is really an acceptable option for the European Commission and ECB (without the necessity of any additional commitment to follow a narrower band) it can give a very broad spectrum of permissible exchange rate mechanisms during the ERM-II trial period: all the way from a currency board to quite a flexible floating regime. Analyzing the advantages and disadvantages of various options, the two extreme solutions in this spectrum seem to be the best choices: either a currency board or a wide fluctuation band. The first one, if backed by a credible political commitment and institutional solutions (i.e. considered by financial market as unchangeable) may cause an immediate convergence of interest rates and make the economic benefits of belonging to the eurozone such as trade creation, financial market integration and investment inflow available earlier. The potential advantages of this variant have been confirmed by the experience of Bulgaria, Estonia and Lithuania (see Chapter 6; Kowalski, 2003; Ganev, Jarocinski, Lubenova & Wozniak, 2001). The second option, i.e. a wide fluctuation band gives a pretty good opportunity to continue a policy of direct inflation targeting, which if credible (as in the case of Czech Republic), well technically operated and backed by a prudent fiscal policy, gives a good chance to meet both the inflation and interest rate criteria without the necessity of defending a declared exchange rate corridor. However, one cannot rule

214

MAREK DABROWSKI

out the risk that financial markets will try to test government and central bank commitment to the established central parity. In addition, financial markets can be uncertain about the final conversion rate of the national currency into the euro. In the case of a currency board the problem of setting the conversion rate is definitely solved at an earlier stage, i.e. during the entry to ERM-II or earlier. The third option, i.e. managing the exchange rate in a narrow band seems to be the worst solution, because it is in conflict with the free movement of capital and the principle of the ‘impossible trinity’ (see Frankel, 1999). This is unable to provide either the credible ex ante exchange rate stability associated with ‘hard’ pegs (because of the risk of speculative pressure and uncertainty about the conversion rate), or the discretion in managing domestic liquidity that comes with free floats. Such regimes may also be less transparent and technically difficult to operate because of a fluctuating demand for money and changing market expectations (see Dabrowski, 2001; 2004). Generally speaking, this kind of exchange regime is prone to speculative attacks, as was experienced by many current EMU members in 19921993 (Wyplosz, 2004)12. Worse, each country participating in the ERM-II must defend the agreed central parity and allowed fluctuation band on its own and cannot count on ECB support, except at the margins, where intervention is compulsory. But reaching the edges of the 15% band is almost certain to be judged as evidence of excessive variability of the exchange rate and of severe tension in its maintenance, and would therefore lead to the Commission and the ECB judging a country to have failed the exchange rate criterion. Surprisingly, the European Commission and ECB have consistently ruled out unilateral euroization as possible variant of ERM-II or more generally, as a transitory solution on the way to full EMU membership, resorting to very unclear economic and legal arguments (see Bratkowski and Rostowski, 2001; Rostowski, 2004). The euro is a fully convertible and internationally tradable currency, so the European Commission and ECB do not have any legal grounds to prohibit any country, which has sufficient international reserves to replace its domestic notes and coins with euros from doing so, whether it belongs to the EU or not. On the other hand, unilateral euroization is not the same as EMU membership, because it neither gives voting rights in the ECB or in the ECOFIN Eurogroup, nor access to ECB refinancing facilities. This means that fears regarding the use of unilateral euroization as a method for circumventing the conditionality of the Maastricht convergence criteria are not justified. However, the European Commission and ECB resistance to the idea of unilateral euroization is the real political fact. Given that full EMU membership seems to be very close for NMS, and currency board arrangements are a very close substitute for unilateral euroization (at least in relation to the ERM-II trial period) the political costs of fighting for the right to unilaterally euroize do not make much sense for NMS. However, it remains an economically attractive option for countries applying for EU membership and having problems with the limited credibility of their monetary policies and currencies. This applies, for example, to the Balkan countries or Turkey, where even currency boards may involve credibility problems, and where a spontaneous euroization is far advanced (see Nuti, 2002; Gros, 2002). Hence, the

A STRATEGY FOR EMU ENLARGEMENT

215

idea of unilateral euroization should not be forgotten, and the objections of the EC and ECB are worth challenging further, on both economic and legal grounds. Another question is connected with the length of the ERM-II trial period. Two years were probably chosen to have enough time to check: (i) correctness of a central parity, which is to become the irrevocable rate of domestic currency conversion into euro; (ii) the sustainability and mutual consistency of macroeconomic policies. However in the case of the EMU founding members, three countries (Austria, Finland and Italy) were formally admitted before the end of the two-year ERM trial period. It is interesting to see whether this precedent will be repeated in relation to NMS. In the cases of Estonia and Lithuania, which have been met all the nominal convergence criteria for a quite long time, thanks to stable and credible currency board mechanisms (Estonia since 1992, Lithuania since 2001) waiting another two years for full EMU membership does not make much sense. On the other hand, the length of the ERM-I trial period did not prevent the current EMU members from nominal divergence once they had adopted the common currency. This increase in inflation dispersion has been caused, to a large extent, by factors which have been beyond the control of national authorities (see Section 12.3.2). This reflects the mutual inconsistency of inflation convergence and exchange rate stability, and raises doubts about rationality of the former as a precondition of adopting a common currency. Fiscal divergence, in turn, reflects flaws of the EU fiscal surveillance mechanism (see Section 12.5.3). 4.4. Setting the conversion rate The conversion rate of a national currency into euro is formally determined only at the end of the ERM trial period. So far, in all cases the central parity of the ERM band was adopted as the conversion rate although in the case of Ireland and Greece central parities were revalued during the ERM trial period. Such an outcome is logical if a country has managed to fulfill the exchange rate stability criterion. Thus, NMS will enjoy a relative freedom in setting their central parities under the ERM-II and, consequently, the rates of conversion. However, this does not imply ease in finding an appropriate conversion rate. Despite several equilibrium exchange rate theories they do not provide a much operational help in setting the concrete nominal conversion rate (see Chapter 7). Thus, a choice of a particular central parity under the ERM-II (with assumption that this will be, most likely, the ultimate conversion rate) will always have an arbitrary character and will involve a risk of error. How big this risk is and what can be consequences of wrong decision? In the case of currency undervaluation a country will have problems with meeting the inflation criterion and financial markets may push authorities to correct this mistake through revaluation of the central parity. With an overvalued currency, the economy may suffer output and employment loses and the risk of payments imbalance, which could force the authorities to devalue central parity and push the country out of the EMU accession track. However, regarding the second danger, the NMS record strong productivity growth (Lenain & Rawdanowicz, 2004), much higher than that

216

MAREK DABROWSKI

of core EMU members, and this creates room for moderate overvaluation of the conversion rate. Balancing various risks, it seems that sticking to the current exchange rate would be the most pragmatic recommendation (see Chapters 6 & 7). This is quite easy in the case of countries running currency boards or stable pegs of their currencies to the euro for a certain period of time. It is less easy task in the case of flexible exchange rate regimes, especially for floaters without any predefined exchange rate trajectory (the Czech Republic, Poland and Slovakia). 4.5. Should the Maastricht criteria be modified? If the criteria of nominal convergence and the ERM-II mechanism have so many flaws and inconsistencies should they perhaps be modified to better reflect the macroeconomic realities of NMS at the beginning of the 21st century? On purely economic grounds the answer is positive at least in the case of two of the criteria: x The inflation criterion should be applied to assess macroeconomic (mostly monetary) policies in the period preceding adoption of the ERM-II central parity and abandoned subsequently. It should be completely abandoned in the case of EMU candidates running well-established euro-denominated currency boards. In addition the reference value should relate to average inflation in the eurozone instead of the average of the three best performing EU members. x The two-year trial period of the exchange rate criterion should be counted ex-post in the case of countries running euro-denominated currency boards or having a stable fixed peg of their national currencies to the euro. Artificial maintenance of the ERM-II requirement (the cases of Estonia and Lithuania) does not have any economic justification. In addition, unilateral euroization should be allowed as one of the variants of ERM-II membership, and EU members should not be discouraged from accelerating their nominal convergence in this way. However, an attempt to modify the Treaty in respect to the nominal convergence criteria and the EMU accession mechanism would probably delay the process of EMU enlargement for many years. The new Treaty establishing a Constitution for Europe has just been signed (on October 29, 2004) and the EU is very unlikely to set up an Intergovernmental Conference to work out any new changes to the Treaty before the current ratification process has been completed. In addition there is the risk of opening a ‘Pandora’s box’: nobody can be sure that the final outcome of the revision process will make the convergence criteria and the EMU accession process more efficient and rational. 4.6. Why is early EMU enlargement beneficial for both candidates and incumbents? Our earlier discussion has pointed to various long-term economic benefits from early adoption of the euro by NMS. In addition, EMU accession, although

A STRATEGY FOR EMU ENLARGEMENT

217

technically not easy and requiring some sacrifices, may have also a very beneficial impact on NMS economies in a short- to medium term perspective. First of all, a clear prospect of rapid entry into the EMU may be a powerful incentive for NMS to discipline macroeconomic policy, particularly in the fiscal sphere. There are two reasons why a later adoption of the euro will involve higher cumulative fiscal costs. First, the postponement of the EMU accession date will be caused in most cases by unwillingness to carry out fiscal consolidation now, which will lead to an accumulation of additional public debt or delay its reduction. Second, later accession will mean higher interest rates during the ‘waiting’ period. Both factors will lead to higher interest payments and, eventually, to a necessity of deeper cuts in non-interest expenditures or higher tax increases when the time of EMU entry comes (Dabrowski, Antczak and Gorzelak, 2004). According to the long term fiscal projections of Gorzelak (2004) the biggest fiscal benefits of early accession can be enjoyed by those countries which at present record the highest debt-to-GDP ratio, the worst primary fiscal balances and the highest interest rates (Poland and Hungary). The result of this simulation is consistent with the experience of Mediterranean countries when they joined the EMU (see Jarocinski, 2003). In addition, fast EMU accession may push politicians in NMS to reform the existing microeconomic rigidities earlier, particularly those relating to labor markets. Although many economists correctly argue for labor market reforms before monetary sovereignty given up, the political economy of policy reforms very often changes the required sequence of actions. Only when politicians lose the monetary and exchange rate instruments for alleviating labor market rigidities (or at least the illusion of having such instruments) do they become ready to take the political risk of unpopular reforms. Such causality has been confirmed by the experience of both the EMU members and those NMS that run currency boards (see Chapter 5). Moreover, delaying EMU accession means not only higher cumulative fiscal costs and later benefits from belonging to the eurozone, but also the risk of nominal divergence and serious macroeconomic and financial turbulence. Looking from the perspective of EMU incumbent members, the benefits of early EMU enlargement also seem to outweigh the risks. First, faster nominal and real convergence means faster economic and social consolidation of the enlarged EU and more effective functioning of the Single European Market. Second, EMU accession by NMS will eliminate the risk of currency crisis on the periphery of the EU. Third, EMU accession of NMS creates an opportunity to speed up their maturing in their new role as MS (instead of continuing ‘applicant’ behavior), increase their involvement in the process of the Union’s deepening, and running responsible macroeconomic policies. Fourth, adoption of the euro by NMS will increase its international importance and international demand for euro. Although the EU has not been very interested in the international expansion of euro, it would be beneficial in many ways for EU economies (for example, partially insulating them from the risk of changes in exchange rates between major currencies).

218

MAREK DABROWSKI 5. EMU AFTER ENLARGEMENT

5.1. Enlargement fears The arguments presented in the previous section should lead to a more active policy of OMS towards EMU Enlargement instead of the ‘don’t rush’ advice, which could be heard until very recently. The ‘don’t rush’ advice of the incumbent side has been probably caused by a number of political and economic concerns. The former relates to a desire to impose on NMS another ‘trial’ period before they become full members of the ‘club’. However, this will mean, an extended period of ‘second class’ membership, and the costs that will involve - limited rights usually mean limited responsibility. One of these economic fears refers to the hypothesis that the accession of fast growing countries will increase inflationary pressure in the eurozone and that, as result, interest rates will have to increase, which would have an additional recessionary impact on the slower growing economies of some of the incumbent members (see Baldwin, Berglof, Giavazzi & Widren, 2001). These fears are mostly unjustified as has been shown in Chapters 1 and 10. The real challenge for slower growing economies, posed by the accession of faster growing countries to EMU, is not inflation but rather making their labor markets sufficiently flexible for them to benefit (rather than loose) from such an expansion of the monetary union. Another fears is the supposed deficit of ‘price-stability culture’ in NMS (see Chapter 1). If true, giving the representatives of NMS full voting rights in the ECB Governing Council (GC) and the ECOFIN Eurogroup, could threaten price stability and macroeconomic prudence in the entire eurozone. Again, this concern does not find support in empirical research, at least in relation to actual inflation performance and expected voting behavior in the ECB GC (see Chapter 11). However, even if a number of small and fast growing economies with an uncertain ‘price stability culture’ joining the EMU was going to change the balance of interests in the ECB GC13, the future voting power of NMS governors has been substantially reduced by the new voting system (based on rotation of voting rights) accepted in 2003 (see below). The same can be said about future voting power of NMS in the ECOFIN Eurogroup in relation to fiscal policy issues. All the NMS but Poland belong to the group of either small or medium-size countries and their overall voting power in any EU decision making bodies is limited (80 votes out of 321 total, according to the Treaty of Nice). Although the biggest NMS (Poland, Czech Republic and Hungary) have the worst fiscal record and might be least willing to back EU measures to enforce fiscal discipline, the same is true for large OMS, with Germany and France permanently in breach of the rules (see Chapter 9; Gros, Mayer & Ubide, 2004). 5.2. The institutional design of ECB governing bodies The recent reform of the ECB voting system14 is officially motivated by the ‘number’ problem, i.e. desire to guarantee an efficient decision making process in the ECB GC when the number of EMU members exceeds 15. Less officially this

A STRATEGY FOR EMU ENLARGEMENT

219

reform was probably motivated by the ‘enlargement fears’ mentioned above by changing the original ‘one-country one-vote’ rule. The ‘number’ problem is not non-existent (see Chapter 11). However, the new solution (the rotation system) will in fact not address this problem. It will limit the number of voting members, but the right of all central bank governors to participate in the ECB GC discussions is retained. On the other hand, voting power inside the ECB GC, will be rebalanced, significantly decreasing the frequency of voting rights of governors from smaller and less developed MS. The criterion for allocating the voting-right frequency (5/6 of which depends on a country’s GDP and 1/6 on the size of its financial sector) will definitely disadvantage NMS and may diminish their interest in fast EMU accession. Indirectly the newly adopted voting pattern assumes that governors represent their countries’ interests and preferences, instead of those of the entire eurozone. Such an assumption is in conflict with the Treaty. How else could one solve the ‘numbers’ problem? The literature suggests at least three other solutions: (1) consolidation of country representations into constituencies; (2) weighted voting; (3) delegating interest rate decision to the ECB Executive Board or to a specially formed monetary policy committee, consisting of individually appointed members (see Chapters 1 & 11 for detail analysis). The first two options, which follow the IMF and World Bank Executive Board voting systems, strengthen the principle of country representation, and this is their main weakness. In addition, option (2) does not solve the ‘numbers’ problem. However, the third option does solve this problem. Furthermore, it guarantees that a panEuropean perspective is adopted in setting ECB interest rates, further strengthens ECB independence, at least in relation to governments and political constituencies in individual MS, and enhances the individual accountability of EB or monetary committee members (see Baldwin et al., 2001). 5.3. The future of EU fiscal surveillance rules Many arguments have been made against the current EU rules for fiscal discipline. The most important are: 1/ The upper limit on fiscal deficits (of 3% of GDP) does not leave room for an active countercyclical fiscal policy during recession or growth slowdown or does not allow the automatic fiscal stabilizers to work. The main weakness of this argument is that it disregards the SGP provision obliging MS to achieve medium-term budgetary positions close to balance or in surplus. The current problems of some EU members in meeting the fiscal deficit requirement, originates from their fiscal laxity in the years of higher economic growth (1999-2001 in the case of the eurozone; see Gros, Mayer and Ubide, 2004). An additional counterargument relates to doubts regarding the short-term effects of fiscal policy. So-called ‘non-Keynesian’ effects can be expected in countries recording high deficit and public debt (see Chapter 5). 2/ The 3% deficit limit neglects public investment needs. Higher investment (and higher deficits) today can contribute to higher rates of economic growth in the

220

MAREK DABROWSKI

future, which will make repaying the debt and lowering the deficit to GDP ratio much easier. This argument is frequently addressed to the situation of NMS, which are expected to have bigger needs in the area of infrastructural investments. The main problem with this argument is that it assumes a substantial positive influence of additional public investment on economic growth and budget revenues. This need not be true. Growth may not accelerate and interest rates may increase as result of higher public sector borrowing, causing debt dynamics to get out of control. Economic history provides many examples of misdirected and ineffective public investment programs. In addition, governments have little incentive to worry about the future solvency of their successors. Their current political needs (for example, forthcoming elections) may push them to over-investment, i.e. to the same kind of fiscal irresponsibility as occurs in the case of other public expenditures. As regards NMS, Gros (2004) argues that their level of public investments is already adequate to their level of per capita income, and they do not need higher deficits for these purposes. 3/ Inconsistency between the fiscal deficit and public debt requirements and greater importance of the latter (see Section 12.4.2 and Chapter 9). While we are sympathetic to giving the debt criterion (and particularly the changes in debt-to-GDP ratio) a bigger importance in the process of monitoring the fiscal positions of MS, we are against abandoning the deficit criterion or its relaxation. On the contrary, the SGP requirement of medium term budget balance or surplus should be given priority, and in the case of countries that have low average growth rates the maximum deficit permitted should be lower than 3% of GDP. What needs to be discussed seriously is the enforcement mechanism of the SGP and Treaty’s fiscal requirements. Although theoretically backed by financial sanctions, these are not automatic. Each step in the excessive deficit procedure requires qualified majority support in the Council of Ministers. In a situation in which an increasing number of MS face problems with meeting the fiscal requirements, a coalition of ‘bad boys’ that will block sanctions may become a frequent phenomenon. At the technical level, the fiscal surveillance mechanism is additionally complicated by the practice of forecasting ex ante a country’s structural (i.e. cyclically adjusted) deficit, which requires one to forecast potential output and other macroeconomic parameters. Experience points to repeated mistakes in forecasting the future growth rates and fiscal positions of individual countries. Furthermore, we do not believe that it is possible to find any formula, which would guarantee the unquestionable (and free of political bargaining) prediction of future structural deficits. In addition, in the case of NMS, which went through a series of deep structural changes during last 15 years, identification of their business cycles and estimation of their potential outputs seems to be a particularly complicated task. As regards improving the enforcement mechanism, Rostowski (Chapter 9) suggests tightening fiscal rules at the national level, preferably in national constitutions (following the example of Poland). Another possibility is to introduce automatic sanctions at the EU level when countries breach the fiscal requirements.

A STRATEGY FOR EMU ENLARGEMENT

221

This could involve, for example, suspending a country’s voting rights in the ECB GC or in the ECOFIN Council of Ministers. Generally we believe that: (i) EU fiscal surveillance rules cannot be relaxed; (ii) their effective implementation requires a much better enforcement mechanism based on automatic sanctions; (iii) NMS do not need different fiscal rules. 6. CONCLUSIONS AND RECOMMENDATIONS The Eastern Enlargement of the EU is not yet complete and will not be complete until NMS join the EMU, a key component of the EU integration package and the Single European Market. If EMU enlargement is delayed beyond the next couple of years, this will mean a significant change in the economic and political architecture of the EU, creating a de facto two-speed Europe. Such a change may in turn cause an anti-integration backlash at the EU periphery, slowing down both nominal and real convergence and prolonging the ‘applicant’ behavior of the NMS (instead of coresponsibility for the entire Union). Both the economic and the political economy arguments reviewed in this paper point, on balance, to fast EMU accession of NMS. First, looking at the ‘classical’ OCA criteria, i.e. trade integration, co-movement of business cycles and actual factor mobility, NMS’ record is not worse, on average, than that of the current eurozone members, and should further improve before eurozone entry, decreasing risk of their exposure to idiosyncratic shocks. Second, after joining the EMU, the common currency should help NMS to develop additional intra-EMU trade links, further synchronize business cycle and increase factor mobility (the ‘endogeneity hypothesis’). These benefits will be bigger, and the risks of asymmetric shocks smaller, if more NMS join the EMU at an earlier stage (so-called network externalities). Third, theoretical arguments and overwhelming empirical experience (including that of Mediterranean members of the EMU) demonstrates that so-called real convergence (understood as catching up with the GDP per capita levels of richer countries) accompanies nominal convergence, and that there is synergy rather than a trade-off between the two. Fourth, a clear perspective of EMU membership can serve as a powerful incentive to conduct politically unpopular fiscal adjustment and microeconomic reforms (for example, making labor markets more flexible). This was observed in Mediterranean countries when they joined the EMU. Fifth, early EMU accession (say, at the beginning of 2007) can provide NMS with significant fiscal gains compared to a later accession date (say, around 2012). OMS, the European Commission and the ECB should replace their ‘don’t rush’ advice by active encouragement of NMS to proceed with fast nominal convergence in order to meet the Maastricht criteria and join the EMU as quickly as possible. Taking into consideration the different fiscal positions of individual NMS, the EMU admission strategy we recommend is a case-by-case approach with immediate accession for countries that have had a hard peg to the euro for many years and meeting the Maastricht fiscal criterion (Estonia and Lithuania). Exposing them to a two-year ERM-II trial period does not make any sense. On the other hand, if both countries were offered a fast track approach, this would create a positive

222

MAREK DABROWSKI

demonstration effect for other EMU candidates and an incentive for them to accelerate their fiscal adjustments. Both the ERM-II mechanism and other nominal convergence criteria designed in the early 1990s (and incorporated into the Maastricht Treaty) suffer a number of flaws and inconsistencies, and do not reflect the new realities of strongly integrated international financial markets dominated by unrestricted capital movements and an already existing eurozone. Thus, it would be good to modify them but such a perspective seems to be politically unrealistic at the moment. As a result, the pace of EMU enlargement will depend very much on the criteria being interpreted reasonably and economically rationally. If used in a formalistic way they may serve to postpone EMU enlargement for many years. Flexibility in interpretation will be needed especially in relation to the exchange rate and inflation criteria and their mutual dependence. If one wants to avoid the danger of speculative pressure during the ERM-II trial period, the best exchange rate mechanisms are either a currency board (first best) or a wide (+/- 15%) fluctuation band (which is second best). A choice of a narrow corridor of +/- 2.25% around the central parity should be discouraged as it would conflict with free movement of capital, and can lead to currency crises. The inflation performance of countries that have chosen the currency board variant, should be assessed with a certain margin of flexibility, on condition that they run a prudent fiscal policy and did not adopt an undervalued exchange rate just before entering the ERM mechanism. In addition, the reference value for the inflation criterion should be the EMU (weighted) average rather than, as at present, the arithmetic average of the three best performing MS (not even necessarily belonging to the EMU). The credibility of the euro and price stability in the eurozone will not be threatened by fast EMU Enlargement. Neither can the accession of fast growing NMS create an additional ‘recessionary’ impact on slow growing incumbent members. The biggest challenge for the common currency in the medium to long run may come from widespread breaches of EU fiscal rules. The domestic political difficulties many MS have with obeying these rules cannot serve as an argument that they are irrelevant for safeguarding fiscal prudence and avoiding fiscal ‘free riding’ under the umbrella of the MU. Growth problems in some OMS and forthcoming population aging in all of Europe, may require even tighter fiscal rules than those currently in the Treaty and SGP. Any fiscal rules (whether the present ones or not) must be solidly anchored in an effective enforcement mechanism (including automatic sanctions) at both the Union and national levels.

NOTES 1

In fact, this assumption has been challenged by some well-respected economists like Feldstein (2000). To make a cross-country analysis of trade links inside the enlarged EU-25 or EU-27 really comprehensive and comparable, one should also add OMS exports to NMS to the trade exposure of the former. However, because of the limited economic potential of NMS this will not substantially increase their intra-EU trade statistics.

2

A STRATEGY FOR EMU ENLARGEMENT

223

3

See e.g. Boone & Maurel (1999), Fidrmuc & Schardax (2000), Korhonen & Fidrmuc (2001), and Gros & Hobza (2003). 4 Twelve OMS have introduced temporary (2 to 7 year) restrictions on labor movement from NMS. 5 In EU terminology ‘cohesion’ has a wider meaning and includes also ‘social cohesion’, which relates to such issues as unemployment, poverty, environmental quality and social exclusion (see Ardy, Begg, Schelke & Torres, 2002, pp. iv & 1). 6 However, from a legal point of view, these are not binding criteria for euro adoption. 7 Surprisingly, labor market integration is not discussed in the Convergence Report. 8 See Sarel (1996) and Gosh & Phillips (1998) for empirical evidence based on a large sample of countries. 9 A similar conclusion has been drawn by Ardy, Begg, Schelke & Torres (2002). 10 The new Treaty establishing a Constitution for Europe signed in Rome on October 29, 2004 and not ratified yet requires, in addition, approval of the new entrant by incumbent EMU members taken by QMV (Article III-198, paragraph 2). 11 The actual (2004) picture does not look so dramatic for NMS. According to CR (2004a, p. 4) four of the NMS-10 satisfy this criterion. Two of them (Estonia and Lithuania) have run currency boards for many years and Cyprus has a stable peg of its pound to euro. The Czech Republic is the only floater meeting this criterion among NMS. 12 A wide band also belongs to the criticized category of intermediate (hybrid) regimes. The main difference is the accepted fluctuation margin, which in the absence of serious shocks and major policy inconsistencies or uncertainties, may allow running a monetary policy close to that with a free float. 13 In fact, these effects will neutralize each other. While a higher growth rate may cause higher inflationary pressure through the HBS effect and push NMS governors (assuming that they look mostly at the monetary conditions in their home country) to vote for higher interest rates in the entire eurozone, a deficit of ‘price stability culture’ would make their position more ‘dovish’. 14 The reform is based on a provision of the Treaty of Nice, and was formulated for the first time by the ECB GC on December 20, 2002. It was accepted by the European Council on March 21, 2003.

REFERENCES Alesina, A., & Perotti, R. (1994). The political economy of budget deficits. IMF Working Paper, WP/94/85. Ardy, B., Begg, I., Schelke, W. & Torres, F. (2002). EMU and its impact on cohesion: policy challenges. EMUCO project report, http://www.lsbu.ac.uk/euroinst/documents/emuco.pdf Baldwin, R.E., Berglof, E., Giavazzi, F. & Widren, M. (2001). Nice try: should the treaty of nice be ratified? Monitoring European Integration, 11, June. London: Centre for Economic Policy Research. Bertola, G. & Drazen, A. (1993). Trigger points and budget cuts: explaining the effects of fiscal austerity. American Economic Review, 83 (1): 11-26. Boone, L. & Maurel, M. (1999). An optimal currency area perspective of the EU enlargement to the CEECs. CEPR Discussion Paper, 2119. London: Centre for Economic Policy Research. Bratkowski, A. & Rostowski, J. (2001). The EU attitude to unilateral euroization: misunderstandings, real concerns and ill-designed admission criteria. [in:] Brzeski A. & Winiecki J. (eds.) A Liberating Economic Journey: Post-communist Transition, Essays in Honour of Ljubo Sirc, London: CRCE. Brzozowski, M. (2003). Exchange rate variability and foreign direct investment – consequences of EMU enlargement. Studies & Analyzes, 258. CASE – Center for Social and Economic Research. CR (2004a). Report from the Commission. Convergence Report 2004. Brussels: Commission of the European Communities, October 20, COM(2004) 690 http://europa.eu.int/comm/economy_finance/publications/european_economy/2004/cr2004_en.pdf CR (2004b). Convergence Report 2004 – Technical annex. A Commission services working paper, SEC(2004) 1268, Brussels: Commission of the European Communities, October 20, http://europa.eu.int/comm/economy_finance/publications/european_economy/convergencereports200 4_en.htm Dabrowski, M. (2001). Global integration of financial markets and its consequences for transition economies. [in:] Orlowski, L.T. (ed.) Transition and Growth in Post-communist countries. The Tenyear Experience, Edgar Elgar, Cheltenham, UK, Northampton, MA, USA.

224

MAREK DABROWSKI

Dabrowski, M. (2004). Costs and benefits of monetary sovereignty in the era of globalization. [in:] Dabrowski, M., Neneman J. & Slay B. (eds.) Beyond Transition. Development Alternatives and Dilemmas. Ashgate Publishing Limited, Aldershot UK, Burlington USA, Sydney Aus. Dabrowski, M. & Rostowski, J., eds. (2001). The eastern enlargement of the EU. Kluwer Academic Publishers, Boston/ Dordrecht/ London. Dabrowski, M., Antczak, M. & Gorzelak, M. (2004). Fiscal challenges facing the new member states. Paper for the DG ECFIN Workshop on ‘Fiscal Surveillance in EMU: New Issues and Challenges’, Brussels, November 12, http://europa.eu.int/comm/economy_finance/events/2004/events_brussels_1104_en.htm ‘Delors Report’ (1989). Report on economic and monetary union in the European community. Committee for the Study of Economic and Monetary Union. Report to European Council by a Committee chaired by J. Delors (the ‘Delors Report’). Office for Official Publications of the European Communities, Luxemburg. Emerson, M., Gros, D., Italianer, A., Pisani-Ferry, J. & Reichenbach, H. (1992). One market, one money: an evaluation of the potential benefits and costs of forming an economic and monetary union. Oxford: Oxford University Press. Feldstein, M. (2000, February 8). Europe can’t handle the euro. Wall Street Journal. Fidrmuc, J. & Schardax, F. (2000). More ‘Pre-Ins’ ante portas? Euro area enlargement, optimum currency area, and nominal convergence. Focus on Transition, 2, 28-47, National Bank of Austria, Vienna. Fischer, S., Sahay R. & Vegh C.A. (1998). How far is eastern Europe from Brussels? IMF Working Paper, WP/98/53, , April. Frankel, J.A. (1999). No single currency regime is right for all countries or at all times. NBER Working Paper, 7338, , September. Frankel, J. A., & Rose, A.K. (1998). The endogeneity of the optimum currency area criteria. Economic Journal, 108, July. Ganev, G., Jarocinski, M., Lubenova, R. & Wozniak, P. (2001). Credibility of the exchange rate policy in transition countries, Reports, 38. CASE – Center for Social and Economic Research. Gomulka, S. (2001). A great leap forward? Pension developments and reforms in the Czech Republic, Hungary, Poland and Romania. [in:] Dabrowski, M. & Rostowski, J. (eds.) The Eastern Enlargement of The EU, Kluwer Academic Publishers, Boston/ Dordrecht/ London. Gorzelak, M. (2004): Fiscal costs of delaying the EMU accession – medium term fiscal projection for selected new EU members. Studies & Analyzes, 277. CASE – Center for Social and Economic Research. Ghosh, A.R. & Phillips, S. (1998). Warning: inflation may be harmful to your growth. IMF Staff Papers, 45 (4). Gros, D. (2002). The euro for the Balkans? Economics of Transition, 10 (2): 491-511. Gros, D. (2004). The fiscal criteria and the new member countries? Paper prepared for CEPS/ESI annual conference on ‘EMU enlargement to the East and West’, Budapest, September 24/25, www.mnb.hu/ Resource.aspx?ResourceID=mnbfile&resourcename=dgros. Gros, D., Mayer, T. & Ubide A. (2004). The nine lives of the stability pact. A Special Report of the CEPS Macroeconomic Policy Group. Center for European Policy Studies, Brussels. Gros, D. & Hobza, A. (2003). Exchange rate variability as an OCA criterion: are the candidates ripe for the euro? Working Papers, 23, Macroeconomic Studies. International Center for Economic Growth European Center. Honohan, P. & Lane, P.R. (2004). Exchange rates and inflation under EMU: An update. CEPR Discussion Paper, 4583, August. London: Centre for Economic Policy Research. Jarocinski, M. (2003). Nominal and real convergence in Spain, Portugal and Greece during their accession to the EMU. Studies and Analysis, 256. CASE – Center for Social and Economic Research. Korhonen, I. & Fidrmuc, J. (2001). Similarity of demand and supply shocks between the Euro Area and the accession countries. Focus on Transition, 2, 26-42, National Bank of Austria, Vienna. Kowalski, P. (2003). Nominal and real convergence in alternative exchange rate regimes in transition countries: implications for the EMU accession. Studies and Analysis, 270. CASE – Center for Social and Economic Research. Lenain, P. & Rawdanowicz, L. (2004). Enhancing income convergence in Central Europe after EU accession. OECD Economics Department Working Papers, 392. OECD Economics Department. Maliszewska, M. (2004). Eastern EU expansion: implications of the enlarged single market for current and new member states. Studies and Analysis, 273. CASE – Center for Social and Economic

A STRATEGY FOR EMU ENLARGEMENT

225

Research. McKinnon, R. (1963). Optimum currency areas. American Economic Review, 53, 717-25. McKinnon, R. (2000). Mundell, the euro and optimum currency areas. Journal of Policy Modeling, 22(3), 311-324. Mundell, R. (1961). A theory of optimum currency areas. American Economic Review, 51, 657- 65. Mundell, R. (1973a). Uncommon arguments for common currencies [in:] Johnson, H. & Swoboda, A. (eds.) The Economics of Common Currencies. London: George Allen & Unwin Ltd, 114-132. Mundell, R. (1973b). A plan for a European currency [in:] Johnson, H. & Swoboda, A. (eds.) The Economics of Common Currencies. London: George Allen & Unwin Ltd , 143-173. Nuti, D.M. (2002). Costs and benefits of unilateral euroization in Central Eastern Europe. Economics of Transition, 10 (2), 419-444. Polackova-Brixi, H. (2004). Contingent liabilities in the new member states. Paper prepared for Workshop: ‘Fiscal Surveillance in EMU: issues and challenges’, Brussels, , November 12, http://europa.eu.int/comm/economy_finance/events/2004/events_brussels_1104_en.htm Reinhart, C.M. & Rogoff, K.S. (2002). The modern history of exchange rate arrangements: A reinterpretation. NBER Working Paper, 8963. Rose, A. (2000). One market, one money: estimating the effect of common currencies on trade. Economic Policy, 30, 7-33, , April. Rostowski, J. (2004). Why unilateral euroization makes sense for (some) applicant countries [in:] Dabrowski, M., Neneman, J. & Slay, B. (eds.) Beyond Transition. Development Alternatives and Dilemmas. Ashgate Publishing Limited, Aldershot UK, Burlington USA, Sydney Aus. Sarel, M. (1996). Nonlinear effects of inflation on economic growth. IMF Staff Papers, 43 (1). Scitovsky, T. (1958). Economic theory and western european integration. Stanford: Stanford University Press. Sutherland, A. (1997). Fiscal crises and aggregate demand: can high public debt reverse the effects of fiscal policy? Journal of Public Economics, 65 (2), 147-162. Vaubel, R. (1976). Real exchange-rate changes in the European Community - The Empirical Evidence and its Implications for European Currency Unification. Weltwirtschaftliches Archiv, 112, 429-470. Vaubel, R. (1978). Real exchange-rate changes in the European Community: A New Approach to the Determination of Optimum Currency Areas. Journal of International Economics, 8, 319-339. von Hagen, J. & Neumann M. (1994). Real exchange rates within and between currency areas: how far away is EMU? The Review of Economics and Statistics. 76 (2), 236-244. Wyplosz, C. (2004). Exchange rate regimes after enlargement [in:] Dabrowski, M., Neneman, J. & Slay, B. (eds.) Beyond Transition. Development Alternatives and Dilemmas. Ashgate Publishing Limited, Aldershot UK, Burlington USA, Sydney Aus.

INDEX

‘numbers’ problem, 219 Accession Treaty, 199 Austria, 3, 5, 14, 43, 44, 45, 47, 49, 60, 68, 70, 148, 195, 204, 215, 224 balance of payments, 8, 16, 17, 41, 111, 118, 202, 203, 207 Balassa-Samuelson effect. See Harrod-Balassa-Samuelson effect bank(s), xi, xii, xiii, 10, 11, 13, 14, 18, 59, 60, 79, 97, 105, 111, 115, 120, 126, 128, 129, 138, 140, 148, 153, 155, 156, 157, 158, 172, 181, 183, 184, 190, 195, 197, 198, 208, 211, 219, 224 banking sector, 18, 97 Belgium, 3, 44, 47, 49, 70, 78, 142, 150, 151, 155, 156, 157, 195 Bulgaria, xi, 3, 43, 44, 45, 47, 49, 51, 52, 53, 55, 58, 59, 60, 100, 101, 102, 138, 205, 213 business (economic) cycle, 41, 46, 48, 58, 125, 150, 156, 157, 199, 202, 203, 204, 221 business (economic) cycle synchronization (co-integration), 201, 202, 203 capital flows, 16, 111, 116, 117, 201, 202 capital inflow, 2, 6, 17, 18, 51, 93, 94, 108, 124 capital mobility, 16, 17, 28, 201 capital outflow, 93, 107 Central and Eastern Europe (CEE), xiii, 14, 73, 110, 111 central bank, xii, 5, 15, 17, 18, 21, 93, 95, 103, 147, 154, 157, 175, 180, 181, 202, 211, 214, 219 Central Europe, 224 central parity, 12, 50, 59, 92, 113, 115, 125, 126, 127, 211, 213, 214, 215, 216, 222 Club Med. See Mediterranean countries cohesion, 207, 210, 223 common currency, 4, 17, 45, 48, 50, 51, 63, 65, 86, 91, 111, 116, 131, 183, 189, 196, 200, 201, 202, 203, 207, 210, 211, 215, 221, 222 constitution, 156, 157 Constitutional Treaty, 196 Consumer Price Index (CPI), 20, 92, 100, 102, 107, 109, 110, 120 convergence, 1, 4, 24, 25, 46, 53, 55, 56, 75, 91, 92, 102, 106, 109, 113, 116, 117, 118, 125, 129, 151, 157, 182, 199, 200, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 221, 222, 224 convergence criteria. See Maastricht criteria convergence nominal, 5, 14 convergence real, xv, 2, 6, 76, 200, 206, 207, 208, 209, 210, 217, 221, 224 Convergence Report, 60, 129, 223 conversion rate, xv, 110, 113, 114, 116, 119, 125, 127, 214, 215, 216 Council of Ministers, 149, 152, 153, 197, 220, 221 countries with derogation, 113, 210, 211 crawling peg (band), 19, 20, 29, 30, 115, 205 currency band, ix, 91, 92, 97, 100, 103, 107, 108, 110, 111 currency board, 1, 29, 30, 51, 52, 79, 82, 92, 94, 98, 100, 102, 108, 109, 114, 115, 204, 205, 206, 208, 213, 214, 215, 222 currency crisis, xi, 2, 42, 80, 95, 128, 217, 222

227

228

INDEX

currency depreciation (devaluation), 77, 201, 202 currency peg, 98, 101, 103 currency union. See monetary union current account balance, 162, 169, 182 Cyprus, 43, 44, 45, 114, 115, 116, 121, 127, 128, 195, 223 Czech Republic, 2, 3, 5, 44, 45, 51, 52, 53, 55, 58, 69, 71, 72, 73, 78, 79, 80, 81, 82, 83, 92, 101, 102, 114, 127, 138, 202, 204, 205, 213, 216, 218, 223, 224 debt, 2, 12, 17, 18, 28, 93, 98, 99, 100, 101, 102, 103, 105, 106, 109, 110, 113, 116, 125, 127, 131, 132, 133, 134, 135, 136, 137, 138, 140, 141, 142, 143, 144, 145, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 209, 210, 211, 212, 217, 219, 220, 225 Denmark, 1, 43, 44, 45, 47, 49, 70, 133, 144, 195, 199, 211 direct inflation targeting, 114, 196, 202, 213 disinflation, 80, 81, 83, 202, 208, 209 dollar, 19, 28, 108, 204, 208 dollarization, 13 Economic and Monetary Union (EMU), v, vii, ix, xi, xv, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 41, 42, 46, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 60, 61, 63, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 78, 79, 80, 81, 82, 83, 84, 85, 86, 91, 92, 93, 94, 97, 98, 99, 100, 101, 102, 104, 105, 106, 108, 111, 113, 114, 115, 116, 117, 119, 123, 124, 125, 126, 127, 128, 148, 151, 152, 154, 155, 156, 158, 159, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 221, 222, 223, 224, 225 Economic and Monetary Union (EMU) enlargement, 42, 63, 113, 115, 126, 148, 183, 189, 196, 199, 200, 206, 208, 211, 216, 217, 218, 221, 222 economic growth, 2, 6, 23, 79, 81, 82, 116, 147, 206, 207, 209, 210, 219, 220, 225 economic reform, 221 emerging markets, 5, 92 Estonia, 1, 3, 5, 12, 14, 43, 45, 46, 47, 49, 51, 52, 53, 55, 58, 59, 69, 70, 71, 72, 73, 78, 102, 113, 114, 115, 127, 138, 195, 205, 213, 215, 216, 221, 223 EU Treaty (Treaty), 6, 12, 51, 106, 149, 150, 154, 155, 157, 183, 184, 192, 193, 196, 200, 206, 207, 210, 211, 216, 218, 219, 220, 222, 223 Euro, v, viii, ix, xv, 13, 14, 44, 45, 49, 50, 51, 52, 53, 56, 59, 60, 90, 94, 101, 102, 103, 108, 109, 110, 111, 113, 114, 116, 119, 121, 122, 123, 124, 126, 127, 128, 140, 185, 197, 198, 224 euroization, 1, 2, 13, 60, 147, 156, 158, 214, 216, 223, 225 Europe, 1, 10, 86, 133, 158, 201, 216, 221, 222, 223, 224, 225 European Central Bank (ECB), v, viii, xv, 7, 9, 10, 11, 13, 51, 53, 55, 56, 57, 60, 98, 127, 129, 148, 155, 156, 157, 183, 184, 186, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 200, 211, 213, 214, 215, 218, 219, 221, 223 European Commission, 1, 13, 14, 59, 60, 98, 151, 153, 154, 158, 191, 200, 207, 213, 214, 221 European Council, 149, 193, 197, 211, 223, 224 European Union (EU), vii, viii, xi, xv, 1, 2, 3, 6, 10, 11, 12, 13, 14, 41, 43, 44, 45, 46, 48, 58, 59, 60, 63, 65, 66, 67, 68, 69, 70, 71, 72, 74, 75, 77, 78, 85, 86, 91, 101, 106, 111, 113, 115, 116, 131, 140, 141, 143, 149, 150, 153, 158, 183, 184, 185, 191, 195, 196, 197, 199, 202, 203, 204, 206, 207, 210, 211, 212, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224 European Union (EU) enlargement, 13, 14

INDEX

229

eurozone (or euroarea), i, vii, xv, 41, 43, 44, 46, 47, 48, 49, 52, 53, 56, 58, 59, 60, 82, 116, 117, 121, 122, 125, 128, 158, 183, 184, 185, 186, 188, 191, 192, 194, 196 exchange rate appreciation, 92, 110, 161, 170, 171, 173, 174 behavioral equilibrium exchange rate (BEER), 117, 118, 119, 121, 124, 126, 127, 128, 129 criterion, 12, 92, 213, 214, 216 damental equilibrium exchange rate (FEER), 117, 118, 119, 123, 124, 128, 129 depreciation, vii, 42, 76, 84, 105 fixed, 2, 16, 17, 18, 41, 81, 91, 92, 93, 100, 106, 109, 120, 122, 127, 208 flexible, 15, 27, 121 floating, 2, 17, 18, 19, 26, 42, 91, 92, 94, 95, 98, 100, 101, 102, 103, 109, 120, 202 fluctuations, 50, 59, 208 misalignment, 119, 124, 126 natural real exchange rate (NATREX), 117, 118, 119, 123, 124, 129 nominal, 7, 17, 75, 77, 114, 127, 161, 162, 169, 170, 171, 174 real, vii, 23, 41, 76, 84, 92, 105, 117, 161, 162, 165, 202 regimes, viii, 13, 28, 110, 114, 115, 225 stability, 92, 94, 97, 101, 109, 205, 208, 211, 212, 213, 214, 215 variability, 50, 54, 59, 63, 64, 74, 203, 204 Exchange Rate Mechanism II (ERM-II) central parity, 127, 216 trial period, 213, 214, 215, 221, 222 expectations, 2, 7, 24, 42, 94, 95, 97, 98, 99, 100, 102, 103, 104, 105, 107, 125, 133, 134, 135, 136, 140, 210, 214 exports, 4, 9, 12, 13, 17, 19, 41, 44, 45, 48, 58, 67, 116, 128, 143, 161, 169, 172, 174, 175, 176, 177, 178, 179, 180, 181, 203, 222 factor mobility, 17, 42, 50, 200, 201, 203, 205, 206, 221 fear of floating, 18, 114 financial markets, 91, 94, 98, 108, 125, 126, 127, 139, 140, 189, 194, 200, 201, 207, 209, 210, 212, 214, 215, 222, 223 Finland, 3, 12, 14, 43, 44, 45, 47, 49, 68, 70, 111, 195, 215 fiscal adjustment, 2, 106, 136, 141, 143, 144, 209, 210, 221 balance, 21, 100, 150, 151, 209 consolidation, 100, 133, 210, 217 counter-cyclical policy, xv criteria, 209, 224 deficit, 6, 12, 17, 59, 134, 149, 150, 151, 219, 220 discipline, 28, 158 discretionary policy, 131 policy, xi, xii, 2, 11, 17, 75, 93, 100, 102, 109, 126, 131, 132, 133, 134, 136, 138, 139, 140, 143, 149, 151, 152, 153, 155, 210, 213, 218, 219, 222 policy with Keynesian effect, 132, 133, 135, 136, 139 policy with non-Keynesian effect, 131, 135, 139, 219 surplus, 25, 149, 152, 219, 220 transfers, 42, 200 foreign direct investments (FDI), 204, 207

230

INDEX

France, 3, 4, 7, 8, 43, 44, 45, 47, 49, 53, 56, 57, 58, 59, 60, 68, 70, 111, 152, 157, 159, 187, 195, 218 fundamentals, ix, 77, 93, 97, 98, 99, 103, 104, 105, 106, 110, 117, 118, 124, 125 Germany, 3, 4, 9, 10, 43, 44, 45, 47, 48, 49, 53, 56, 57, 59, 60, 68, 70, 74, 148, 149, 152, 157, 159, 187, 188, 195, 218 Greece, 3, 12, 43, 44, 45, 46, 47, 48, 49, 56, 57, 58, 70, 92, 150, 151, 155, 156, 195, 199, 208, 209, 215, 224 Gross Domestic Product (GDP), vii, ix, 3, 4, 5, 6, 7, 11, 12, 13, 19, 20, 21, 23, 24, 25, 26, 27, 30, 31, 37, 38, 39, 40, 43, 44, 45, 46, 48, 58, 59, 66, 67, 68, 69, 70, 71, 79, 80, 81, 83, 87, 100, 101, 106, 110, 127, 128, 138, 139, 143, 144, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 169, 193, 195, 197, 203, 211, 212, 217, 219, 220, 221 hard peg, 17, 221 Harrod-Balassa-Samuelson (HBS) effect, 7, 8, 94, 110, 128, 159, 161, 163, 165, 166, 167, 169, 171, 173, 212, 223 Hungary, xiii, 2, 3, 5, 12, 13, 43, 45, 46, 47, 48, 49, 51, 52, 55, 58, 59, 69, 70, 71, 72, 73, 78, 80, 81, 83, 84, 102, 109, 111, 114, 115, 116, 126, 127, 138, 195, 203, 205, 209, 217, 218, 224 imports, 16, 19, 44, 67, 78, 110, 143, 161, 169, 174 income, xi, xiii, 4, 6, 9, 13, 24, 25, 26, 63, 69, 71, 72, 73, 131, 132, 133, 134, 135, 136, 137, 138, 142, 161, 165, 166, 168, 169, 171, 174, 175, 178, 180, 181, 207, 208, 209, 212, 220, 224 inflation convergence, 182, 215 criterion, 7, 91, 92, 94, 212, 215, 216, 222 differences, 212 interest payments, 209, 217 interest rate convergence, 208, 209, 213 criterion, 12, 93, 208, 213 differentials, 124 nominal, 211 real, 6, 8 international reserves, xiii, 81, 118, 201, 214 Ireland, xi, 3, 7, 43, 44, 45, 47, 49, 70, 92, 133, 144, 152, 157, 195, 208, 215 Italy, 3, 4, 5, 7, 8, 43, 44, 45, 47, 49, 56, 57, 70, 136, 145, 149, 150, 151, 155, 156, 157, 159, 187, 195, 208, 209, 215 labor market flexibility, 10, 75, 77, 78 reform, 181, 217 regulation(s), 100 rigidities, xv, 181, 204, 206, 217 Latvia, xi, 3, 12, 43, 44, 45, 46, 47, 49, 50, 51, 52, 53, 55, 58, 59, 69, 70, 71, 72, 73, 78, 81, 83, 84, 109, 114, 115, 127, 138, 195, 204, 205 legal criteria, 207, 211 lender of last resort, 18, 155 Lithuania, 1, 3, 5, 12, 43, 44, 45, 46, 47, 49, 51, 52, 53, 55, 58, 60, 69, 70, 71, 72, 73, 78, 82, 83, 84, 85, 102, 108, 109, 113, 114, 115, 127, 138, 195, 204, 205, 206, 213, 215, 216, 221, 223

INDEX

231

Luxemburg, 195, 224 Maastricht criteria, 2, 10, 11, 91, 94, 101, 102, 125, 126, 127, 150, 199, 206, 208, 216, 221 Maastricht Treaty, 149, 209, 222 Macroeconomic Policy, xi, 224 macroeconomic stability, 1, 158, 207 Malta, 114, 115, 121, 127, 195 Mediterranean countries, 205, 217, 221 Member States new member states (NMS), vii, viii, xv, 1, 2, 3, 4, 5, 6, 7, 10, 11, 12, 41, 43, 44, 45, 46, 47, 48, 50, 51, 52, 53, 55, 56, 57, 58, 59, 63, 67, 69, 70, 71, 72, 73, 75, 76, 78, 79, 84, 85, 91, 92, 94, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 113, 114, 115, 116, 117, 119, 120, 121, 122, 123, 125, 126, 127, 128, 182, 183, 185, 189, 195, 196, 199, 200, 202, 203, 204, 205, 206, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223 old member states (OMS), vii, 44, 45, 46, 48, 58, 69, 70, 71, 72, 73, 92, 98, 183, 195, 199, 200, 203, 206, 208, 212, 218, 221, 222, 223 with derogation, 210 monetary authority(ies), 18, 105, 174, 181, 212 Monetary Policy, xi, 111, 197, 198 monetization, 100 money demand, 21, 95, 96, 97, 106, 107, 110 supply, 16, 17, 50, 95, 96, 98, 106, 107, 170 Netherlands, 3, 43, 44, 45, 47, 49, 70, 148, 195, 208 OECD, viii, xii, 56, 57, 121, 123, 127, 128, 131, 132, 136, 137, 138, 140, 141, 142, 143, 144, 145, 224 optimum currency area (OCA), 1, 3, 4, 12, 14, 17, 41, 42, 43, 44, 46, 48, 49, 50, 52, 54, 58, 60, 75, 76, 81, 85, 86, 200, 201, 203, 206, 207, 210, 221, 224 opt-out option, 211 permanent income life cycle hypothesis (PILCH), 131, 132, 133, 135, 136, 139, 140 Poland, xi, xiii, xv, 2, 3, 5, 43, 44, 45, 46, 47, 48, 49, 51, 52, 53, 55, 58, 60, 69, 70, 71, 72, 73, 78, 81, 82, 83, 84, 86, 101, 114, 115, 127, 138, 154, 190, 195, 202, 204, 205, 209, 216, 217, 218, 220, 224 population aging, 212, 222 Portugal, 3, 43, 44, 45, 47, 48, 49, 56, 57, 58, 70, 126, 151, 157, 195, 199, 208, 224 productivity growth, xv, 8, 13, 76, 80, 82, 83, 84, 85, 99, 110, 159, 160, 161, 162, 163, 165, 169, 170, 171, 172, 174, 176, 177, 179, 180, 181, 182, 189, 206, 209, 212, 215 purchasing power parity (PPP), viii, 106, 113, 118, 119, 122, 123, 125, 126, 128, 129, 212 qualified majority voting (QMV), 191, 197, 223 Ricardian equivalence, 132, 133, 140 risk premium, 6 Romania, xi, xii, 3, 12, 43, 44, 45, 47, 49, 51, 52, 53, 55, 58, 59, 60, 100, 101, 102, 109, 110, 138, 195, 205, 224 seigniorage, 147, 155, 157 shocks asymmetric (idiosyncratic), 3, 22, 47, 50, 201, 203, 206, 221 demand, 4, 5 nominal, 16, 52, 58, 203, 205

232

INDEX

real, 26, 52, 58, 59, 205 supply, 5, 14, 224 Single European Market, 70, 205, 217, 221 Slovakia, 3, 5, 12, 43, 44, 45, 46, 47, 49, 51, 52, 55, 58, 69, 70, 71, 72, 73, 78, 81, 82, 83, 84, 114, 115, 127, 128, 138, 188, 195, 204, 205, 216 Slovenia, 3, 5, 43, 45, 46, 47, 48, 49, 51, 52, 55, 58, 59, 69, 70, 71, 72, 73, 78, 100, 102, 113, 114, 115, 127, 195, 203, 204, 205, 213 Spain, 3, 4, 7, 43, 44, 45, 47, 49, 56, 57, 68, 70, 92, 126, 151, 159, 195, 199, 208, 224 Stability and Growth Pact (SGP), xv, 6, 11, 13, 137, 138, 139, 141, 147, 149, 150, 151, 152, 153, 154, 206, 209, 210, 219, 220, 222 Sweden, 43, 44, 45, 47, 49, 68, 70, 190, 195, 199 trade creation, xv, 201, 213 integration, vii, 3, 5, 71, 203, 204, 206, 221 transaction costs, 16, 63, 66, 68, 70, 199, 200, 203 transition countries (economies), viii, xii, 46, 47, 86, 102, 129, 131, 132, 136, 138, 139, 140, 145, 204, 208, 209, 223, 224 transparency, 10, 194 uncovered interest rate parity, 128 unemployment, 5, 9, 10, 42, 46, 47, 48, 49, 58, 75, 77, 78, 79, 80, 81, 82, 83, 84, 85, 97, 168, 173, 175, 178, 180, 181, 182, 185, 203, 204, 223 United Kingdom (UK), xiii, 12, 43, 44, 45, 47, 49, 78, 113, 151, 152, 157, 187, 190, 192, 195, 196, 197, 199, 211, 223, 224, 225 wage(s) flexibility, 75, 76, 79 nominal, 75, 76, 77, 78, 80, 81, 82, 83, 85, 173, 174, 178, 180 real, 78, 79, 80, 81, 82, 83, 84, 85, 173, 174, 175, 176, 179 rigidity, 77, 79, 84, 174, 175, 180 stickiness, 76